Merge branch 'main' into memory_usage3

Co-authored-by: gemini-cli-robot <224641728+gemini-cli-robot@users.noreply.github.com>
Co-authored-by: Sam Roberts <158088236+g-samroberts@users.noreply.github.com>

.github/workflows/agent-session-drift-check.yml

@@ -0,0 +1,132 @@
+# yaml-language-server: $schema=https://json.schemastore.org/github-workflow.json
+
+name: 'Agent Session Drift Check'
+
+on:
+  pull_request:
+    branches:
+      - 'main'
+      - 'release/**'
+    paths:
+      - 'packages/cli/src/nonInteractiveCli.ts'
+      - 'packages/cli/src/nonInteractiveCliAgentSession.ts'
+
+concurrency:
+  group: '${{ github.workflow }}-${{ github.head_ref || github.ref }}'
+  cancel-in-progress: true
+
+jobs:
+  check-drift:
+    name: 'Check Agent Session Drift'
+    runs-on: 'ubuntu-latest'
+    if: "github.repository == 'google-gemini/gemini-cli'"
+    permissions:
+      contents: 'read'
+      pull-requests: 'write'
+    steps:
+      - name: 'Detect drift and comment'
+        uses: 'actions/github-script@60a0d83039c74a4aee543508d2ffcb1c3799cdea' # ratchet:actions/github-script@v8
+        with:
+          script: |-
+            // === Pair configuration — append here to cover more pairs ===
+            const PAIRS = [
+              {
+                legacy: 'packages/cli/src/nonInteractiveCli.ts',
+                session: 'packages/cli/src/nonInteractiveCliAgentSession.ts',
+                label: 'non-interactive CLI',
+              },
+              // Future pairs can be added here. Remember to also add both
+              // paths to the `paths:` filter at the top of this workflow.
+              // Example:
+              // {
+              //   legacy: 'packages/core/src/agents/local-invocation.ts',
+              //   session: 'packages/core/src/agents/local-session-invocation.ts',
+              //   label: 'local subagent invocation',
+              // },
+            ];
+            // ============================================================
+
+            const prNumber = context.payload.pull_request.number;
+            const { owner, repo } = context.repo;
+
+            // Use the API to list changed files — no checkout/git diff needed.
+            const files = await github.paginate(github.rest.pulls.listFiles, {
+              owner,
+              repo,
+              pull_number: prNumber,
+              per_page: 100,
+            });
+            const changed = new Set(files.map((f) => f.filename));
+
+            const warnings = [];
+            for (const { legacy, session, label } of PAIRS) {
+              const legacyChanged = changed.has(legacy);
+              const sessionChanged = changed.has(session);
+              if (legacyChanged && !sessionChanged) {
+                warnings.push(
+                  `**${label}**: \`${legacy}\` was modified but \`${session}\` was not.`,
+                );
+              } else if (!legacyChanged && sessionChanged) {
+                warnings.push(
+                  `**${label}**: \`${session}\` was modified but \`${legacy}\` was not.`,
+                );
+              }
+            }
+
+            const MARKER = '<!-- agent-session-drift-check -->';
+
+            // Look up our existing drift comment (for upsert/cleanup).
+            const comments = await github.paginate(github.rest.issues.listComments, {
+              owner,
+              repo,
+              issue_number: prNumber,
+              per_page: 100,
+            });
+            const existing = comments.find(
+              (c) => c.user?.type === 'Bot' && c.body?.includes(MARKER),
+            );
+
+            if (warnings.length === 0) {
+              core.info('No drift detected.');
+              // If drift was previously flagged and is now resolved, remove the comment.
+              if (existing) {
+                await github.rest.issues.deleteComment({
+                  owner,
+                  repo,
+                  comment_id: existing.id,
+                });
+                core.info(`Deleted stale drift comment ${existing.id}.`);
+              }
+              return;
+            }
+
+            const body = [
+              MARKER,
+              '### ⚠️ Invocation Drift Warning',
+              '',
+              'The following file pairs should generally be kept in sync during the AgentSession migration:',
+              '',
+              ...warnings.map((w) => `- ${w}`),
+              '',
+              'If this is intentional (e.g., a bug fix specific to one implementation), you can ignore this comment.',
+              '',
+              '_This check will be removed once the legacy implementations are deleted._',
+            ].join('\n');
+
+            if (existing) {
+              core.info(`Updating existing drift comment ${existing.id}.`);
+              await github.rest.issues.updateComment({
+                owner,
+                repo,
+                comment_id: existing.id,
+                body,
+              });
+            } else {
+              core.info('Creating new drift comment.');
+              await github.rest.issues.createComment({
+                owner,
+                repo,
+                issue_number: prNumber,
+                body,
+              });
+            }

.github/workflows/chained_e2e.yml

@@ -183,7 +183,7 @@ jobs:
     needs:
       - 'merge_queue_skipper'
       - 'parse_run_context'
-    runs-on: 'macos-latest'
+    runs-on: 'macos-latest-large'
     if: |
       github.repository == 'google-gemini/gemini-cli' && always() && (needs.merge_queue_skipper.result !='success' || needs.merge_queue_skipper.outputs.skip != 'true')
     steps:

.github/workflows/ci.yml

@@ -224,7 +224,7 @@ jobs:
 
   test_mac:
     name: 'Test (Mac) - ${{ matrix.node-version }}, ${{ matrix.shard }}'
-    runs-on: 'macos-latest'
+    runs-on: 'macos-latest-large'
     needs:
       - 'merge_queue_skipper'
     if: "github.repository == 'google-gemini/gemini-cli' && needs.merge_queue_skipper.outputs.skip == 'false'"

.github/workflows/deflake.yml

@@ -77,7 +77,7 @@ jobs:
 
   deflake_e2e_mac:
     name: 'E2E Test (macOS)'
-    runs-on: 'macos-latest'
+    runs-on: 'macos-latest-large'
     if: "github.repository == 'google-gemini/gemini-cli'"
     steps:
       - name: 'Checkout'

docs/changelogs/index.md

@@ -18,6 +18,27 @@ on GitHub.
 | [Preview](preview.md) | Experimental features ready for early feedback. |
 | [Stable](latest.md)   | Stable, recommended for general use.            |
 
+## Announcements: v0.38.0 - 2026-04-14
+
+- **Chapters Narrative Flow:** Group agent interactions into "Chapters" based on
+  intent and tool usage for better session structure
+  ([#23150](https://github.com/google-gemini/gemini-cli/pull/23150) by
+  @Abhijit-2592,
+  [#24079](https://github.com/google-gemini/gemini-cli/pull/24079) by
+  @gundermanc).
+- **Context Compression Service:** Advanced context management to efficiently
+  distill conversation history
+  ([#24483](https://github.com/google-gemini/gemini-cli/pull/24483) by
+  @joshualitt).
+- **UI Flicker & UX Enhancements:** Solved rendering flicker with "Terminal
+  Buffer" mode and introduced selective topic expansion
+  ([#24512](https://github.com/google-gemini/gemini-cli/pull/24512) by
+  @jacob314, [#24793](https://github.com/google-gemini/gemini-cli/pull/24793) by
+  @Abhijit-2592).
+- **Persistent Policy Approvals:** Implemented context-aware persistent
+  approvals for tool execution
+  ([#23257](https://github.com/google-gemini/gemini-cli/pull/23257) by @jerop).
+
 ## Announcements: v0.37.0 - 2026-04-08
 
 - **Dynamic Sandbox Expansion:** Implemented dynamic sandbox expansion and

docs/changelogs/latest.md

@@ -1,6 +1,6 @@
-# Latest stable release: v0.37.1
+# Latest stable release: v0.38.1
 
-Released: April 09, 2026
+Released: April 15, 2026
 
 For most users, our latest stable release is the recommended release. Install
 the latest stable version with:
@@ -11,415 +11,261 @@ npm install -g @google/gemini-cli
 
 ## Highlights
 
-- **Dynamic Sandbox Expansion:** Implemented dynamic sandbox expansion and
-  worktree support for both Linux and Windows, enhancing development flexibility
-  in restricted environments.
-- **Tool-Based Topic Grouping (Chapters):** Introduced "Chapters" to logically
-  group agent interactions based on tool usage and intent, providing a clearer
-  narrative flow in long sessions.
-- **Enhanced Browser Agent:** Added persistent session management, dynamic
-  read-only tool discovery, and sandbox-aware initialization for the browser
-  agent.
-- **Security & Permission Hardening:** Implemented secret visibility lockdown
-  for environment files and integrated integrity controls for Windows
-  sandboxing.
+- **Chapters Narrative Flow:** Introduced tool-based topic grouping ("Chapters")
+  to provide better session structure and narrative continuity in long-running
+  tasks.
+- **Context Compression Service:** Implemented a dedicated service for advanced
+  context management, efficiently distilling conversation history to preserve
+  focus and tokens.
+- **Enhanced UI Stability & UX:** Introduced a new "Terminal Buffer" mode to
+  solve rendering flicker, along with selective topic expansion and improved
+  tool confirmation layouts.
+- **Context-Aware Policy Approvals:** Users can now grant persistent,
+  context-aware approvals for tools, significantly reducing manual confirmation
+  overhead for trusted workflows.
+- **Background Process Monitoring:** New tools for monitoring and inspecting
+  background shell processes, providing better visibility into asynchronous
+  tasks.
 
 ## What's Changed
 
-- fix(acp): handle all InvalidStreamError types gracefully in prompt
-  [#24540](https://github.com/google-gemini/gemini-cli/pull/24540)
-- feat(acp): add support for /about command
-  [#24649](https://github.com/google-gemini/gemini-cli/pull/24649)
-- feat(acp): add /help command
-  [#24839](https://github.com/google-gemini/gemini-cli/pull/24839)
-- feat(evals): centralize test agents into test-utils for reuse by @Samee24 in
-  [#23616](https://github.com/google-gemini/gemini-cli/pull/23616)
-- revert: chore(config): disable agents by default by @abhipatel12 in
-  [#23672](https://github.com/google-gemini/gemini-cli/pull/23672)
-- fix(plan): update telemetry attribute keys and add timestamp by @Adib234 in
-  [#23685](https://github.com/google-gemini/gemini-cli/pull/23685)
-- fix(core): prevent premature MCP discovery completion by @jackwotherspoon in
-  [#23637](https://github.com/google-gemini/gemini-cli/pull/23637)
-- feat(browser): add maxActionsPerTask for browser agent setting by
-  @cynthialong0-0 in
-  [#23216](https://github.com/google-gemini/gemini-cli/pull/23216)
-- fix(core): improve agent loader error formatting for empty paths by
-  @adamfweidman in
-  [#23690](https://github.com/google-gemini/gemini-cli/pull/23690)
-- fix(cli): only show updating spinner when auto-update is in progress by
-  @scidomino in [#23709](https://github.com/google-gemini/gemini-cli/pull/23709)
-- Refine onboarding metrics to log the duration explicitly and use the tier
-  name. by @yunaseoul in
-  [#23678](https://github.com/google-gemini/gemini-cli/pull/23678)
-- chore(tools): add toJSON to tools and invocations to reduce logging verbosity
-  by @alisa-alisa in
-  [#22899](https://github.com/google-gemini/gemini-cli/pull/22899)
-- fix(cli): stabilize copy mode to prevent flickering and cursor resets by
-  @mattKorwel in
-  [#22584](https://github.com/google-gemini/gemini-cli/pull/22584)
-- fix(test): move flaky ctrl-c-exit test to non-blocking suite by @mattKorwel in
-  [#23732](https://github.com/google-gemini/gemini-cli/pull/23732)
-- feat(skills): add ci skill for automated failure replication by @mattKorwel in
-  [#23720](https://github.com/google-gemini/gemini-cli/pull/23720)
-- feat(sandbox): implement forbiddenPaths for OS-specific sandbox managers by
-  @ehedlund in [#23282](https://github.com/google-gemini/gemini-cli/pull/23282)
-- fix(core): conditionally expose additional_permissions in shell tool by
-  @galz10 in [#23729](https://github.com/google-gemini/gemini-cli/pull/23729)
-- refactor(core): standardize OS-specific sandbox tests and extract linux helper
-  methods by @ehedlund in
-  [#23715](https://github.com/google-gemini/gemini-cli/pull/23715)
-- format recently added script by @scidomino in
-  [#23739](https://github.com/google-gemini/gemini-cli/pull/23739)
-- fix(ui): prevent over-eager slash subcommand completion by @keithguerin in
-  [#20136](https://github.com/google-gemini/gemini-cli/pull/20136)
-- Fix dynamic model routing for gemini 3.1 pro to customtools model by
-  @kevinjwang1 in
-  [#23641](https://github.com/google-gemini/gemini-cli/pull/23641)
-- feat(core): support inline agentCardJson for remote agents by @adamfweidman in
-  [#23743](https://github.com/google-gemini/gemini-cli/pull/23743)
-- fix(cli): skip console log/info in headless mode by @cynthialong0-0 in
-  [#22739](https://github.com/google-gemini/gemini-cli/pull/22739)
-- test(core): install bubblewrap on Linux CI for sandbox integration tests by
-  @ehedlund in [#23583](https://github.com/google-gemini/gemini-cli/pull/23583)
-- docs(reference): split tools table into category sections by @sheikhlimon in
-  [#21516](https://github.com/google-gemini/gemini-cli/pull/21516)
-- fix(browser): detect embedded URLs in query params to prevent allowedDomains
-  bypass by @tony-shi in
-  [#23225](https://github.com/google-gemini/gemini-cli/pull/23225)
-- fix(browser): add proxy bypass constraint to domain restriction system prompt
-  by @tony-shi in
-  [#23229](https://github.com/google-gemini/gemini-cli/pull/23229)
-- fix(policy): relax write_file argsPattern in plan mode to allow paths without
-  session ID by @Adib234 in
-  [#23695](https://github.com/google-gemini/gemini-cli/pull/23695)
-- docs: fix grammar in CONTRIBUTING and numbering in sandbox docs by
-  @splint-disk-8i in
-  [#23448](https://github.com/google-gemini/gemini-cli/pull/23448)
-- fix(acp): allow attachments by adding a permission prompt by @sripasg in
-  [#23680](https://github.com/google-gemini/gemini-cli/pull/23680)
-- fix(core): thread AbortSignal to chat compression requests (#20405) by
-  @SH20RAJ in [#20778](https://github.com/google-gemini/gemini-cli/pull/20778)
-- feat(core): implement Windows sandbox dynamic expansion Phase 1 and 2.1 by
-  @scidomino in [#23691](https://github.com/google-gemini/gemini-cli/pull/23691)
-- Add note about root privileges in sandbox docs by @diodesign in
-  [#23314](https://github.com/google-gemini/gemini-cli/pull/23314)
-- docs(core): document agent_card_json string literal options for remote agents
-  by @adamfweidman in
-  [#23797](https://github.com/google-gemini/gemini-cli/pull/23797)
-- fix(cli): resolve TTY hang on headless environments by unconditionally
-  resuming process.stdin before React Ink launch by @cocosheng-g in
-  [#23673](https://github.com/google-gemini/gemini-cli/pull/23673)
-- fix(ui): cleanup estimated string length hacks in composer by @keithguerin in
-  [#23694](https://github.com/google-gemini/gemini-cli/pull/23694)
-- feat(browser): dynamically discover read-only tools by @cynthialong0-0 in
-  [#23805](https://github.com/google-gemini/gemini-cli/pull/23805)
-- docs: clarify policy requirement for `general.plan.directory` in settings
-  schema by @jerop in
-  [#23784](https://github.com/google-gemini/gemini-cli/pull/23784)
-- Revert "perf(cli): optimize --version startup time (#23671)" by @scidomino in
-  [#23812](https://github.com/google-gemini/gemini-cli/pull/23812)
-- don't silence errors from wombat by @scidomino in
-  [#23822](https://github.com/google-gemini/gemini-cli/pull/23822)
-- fix(ui): prevent escape key from cancelling requests in shell mode by
-  @PrasannaPal21 in
-  [#21245](https://github.com/google-gemini/gemini-cli/pull/21245)
-- Changelog for v0.36.0-preview.0 by @gemini-cli-robot in
-  [#23702](https://github.com/google-gemini/gemini-cli/pull/23702)
-- feat(core,ui): Add experiment-gated support for gemini flash 3.1 lite by
-  @chrstnb in [#23794](https://github.com/google-gemini/gemini-cli/pull/23794)
-- Changelog for v0.36.0-preview.3 by @gemini-cli-robot in
-  [#23827](https://github.com/google-gemini/gemini-cli/pull/23827)
-- new linting check: github-actions-pinning by @alisa-alisa in
-  [#23808](https://github.com/google-gemini/gemini-cli/pull/23808)
-- fix(cli): show helpful guidance when no skills are available by @Niralisj in
-  [#23785](https://github.com/google-gemini/gemini-cli/pull/23785)
-- fix: Chat logs and errors handle tail tool calls correctly by @googlestrobe in
-  [#22460](https://github.com/google-gemini/gemini-cli/pull/22460)
-- Don't try removing a tag from a non-existent release. by @scidomino in
-  [#23830](https://github.com/google-gemini/gemini-cli/pull/23830)
-- fix(cli): allow ask question dialog to take full window height by @jacob314 in
-  [#23693](https://github.com/google-gemini/gemini-cli/pull/23693)
-- fix(core): strip leading underscores from error types in telemetry by
-  @yunaseoul in [#23824](https://github.com/google-gemini/gemini-cli/pull/23824)
-- Changelog for v0.35.0 by @gemini-cli-robot in
-  [#23819](https://github.com/google-gemini/gemini-cli/pull/23819)
-- feat(evals): add reliability harvester and 500/503 retry support by
-  @alisa-alisa in
-  [#23626](https://github.com/google-gemini/gemini-cli/pull/23626)
-- feat(sandbox): dynamic Linux sandbox expansion and worktree support by @galz10
-  in [#23692](https://github.com/google-gemini/gemini-cli/pull/23692)
-- Merge examples of use into quickstart documentation by @diodesign in
-  [#23319](https://github.com/google-gemini/gemini-cli/pull/23319)
-- fix(cli): prioritize primary name matches in slash command search by @sehoon38
-  in [#23850](https://github.com/google-gemini/gemini-cli/pull/23850)
-- Changelog for v0.35.1 by @gemini-cli-robot in
-  [#23840](https://github.com/google-gemini/gemini-cli/pull/23840)
-- fix(browser): keep input blocker active across navigations by @kunal-10-cloud
-  in [#22562](https://github.com/google-gemini/gemini-cli/pull/22562)
-- feat(core): new skill to look for duplicated code while reviewing PRs by
-  @devr0306 in [#23704](https://github.com/google-gemini/gemini-cli/pull/23704)
-- fix(core): replace hardcoded non-interactive ASK_USER denial with explicit
-  policy rules by @ruomengz in
-  [#23668](https://github.com/google-gemini/gemini-cli/pull/23668)
-- fix(plan): after exiting plan mode switches model to a flash model by @Adib234
-  in [#23885](https://github.com/google-gemini/gemini-cli/pull/23885)
-- feat(gcp): add development worker infrastructure by @mattKorwel in
-  [#23814](https://github.com/google-gemini/gemini-cli/pull/23814)
-- fix(a2a-server): A2A server should execute ask policies in interactive mode by
-  @kschaab in [#23831](https://github.com/google-gemini/gemini-cli/pull/23831)
-- feat(core): define TrajectoryProvider interface by @sehoon38 in
-  [#23050](https://github.com/google-gemini/gemini-cli/pull/23050)
-- Docs: Update quotas and pricing by @jkcinouye in
-  [#23835](https://github.com/google-gemini/gemini-cli/pull/23835)
-- fix(core): allow disabling environment variable redaction by @galz10 in
-  [#23927](https://github.com/google-gemini/gemini-cli/pull/23927)
-- feat(cli): enable notifications cross-platform via terminal bell fallback by
-  @genneth in [#21618](https://github.com/google-gemini/gemini-cli/pull/21618)
-- feat(sandbox): implement secret visibility lockdown for env files by
-  @DavidAPierce in
-  [#23712](https://github.com/google-gemini/gemini-cli/pull/23712)
-- fix(core): remove shell outputChunks buffer caching to prevent memory bloat
-  and sanitize prompt input by @spencer426 in
-  [#23751](https://github.com/google-gemini/gemini-cli/pull/23751)
-- feat(core): implement persistent browser session management by @kunal-10-cloud
-  in [#21306](https://github.com/google-gemini/gemini-cli/pull/21306)
-- refactor(core): delegate sandbox denial parsing to SandboxManager by
-  @scidomino in [#23928](https://github.com/google-gemini/gemini-cli/pull/23928)
-- dep(update) Update Ink version to 6.5.0 by @jacob314 in
-  [#23843](https://github.com/google-gemini/gemini-cli/pull/23843)
-- Docs: Update 'docs-writer' skill for relative links by @jkcinouye in
-  [#21463](https://github.com/google-gemini/gemini-cli/pull/21463)
-- Changelog for v0.36.0-preview.4 by @gemini-cli-robot in
-  [#23935](https://github.com/google-gemini/gemini-cli/pull/23935)
-- fix(acp): Update allow approval policy flow for ACP clients to fix config
-  persistence and compatible with TUI by @sripasg in
-  [#23818](https://github.com/google-gemini/gemini-cli/pull/23818)
-- Changelog for v0.35.2 by @gemini-cli-robot in
-  [#23960](https://github.com/google-gemini/gemini-cli/pull/23960)
-- ACP integration documents by @g-samroberts in
-  [#22254](https://github.com/google-gemini/gemini-cli/pull/22254)
-- fix(core): explicitly set error names to avoid bundling renaming issues by
-  @yunaseoul in [#23913](https://github.com/google-gemini/gemini-cli/pull/23913)
-- feat(core): subagent isolation and cleanup hardening by @abhipatel12 in
-  [#23903](https://github.com/google-gemini/gemini-cli/pull/23903)
-- disable extension-reload test by @scidomino in
-  [#24018](https://github.com/google-gemini/gemini-cli/pull/24018)
-- feat(core): add forbiddenPaths to GlobalSandboxOptions and refactor
-  createSandboxManager by @ehedlund in
-  [#23936](https://github.com/google-gemini/gemini-cli/pull/23936)
-- refactor(core): improve ignore resolution and fix directory-matching bug by
-  @ehedlund in [#23816](https://github.com/google-gemini/gemini-cli/pull/23816)
-- revert(core): support custom base URL via env vars by @spencer426 in
-  [#23976](https://github.com/google-gemini/gemini-cli/pull/23976)
-- Increase memory limited for eslint. by @jacob314 in
-  [#24022](https://github.com/google-gemini/gemini-cli/pull/24022)
-- fix(acp): prevent crash on empty response in ACP mode by @sripasg in
-  [#23952](https://github.com/google-gemini/gemini-cli/pull/23952)
-- feat(core): Land `AgentHistoryProvider`. by @joshualitt in
-  [#23978](https://github.com/google-gemini/gemini-cli/pull/23978)
-- fix(core): switch to subshells for shell tool wrapping to fix heredocs and
-  edge cases by @abhipatel12 in
-  [#24024](https://github.com/google-gemini/gemini-cli/pull/24024)
-- Debug command. by @jacob314 in
-  [#23851](https://github.com/google-gemini/gemini-cli/pull/23851)
-- Changelog for v0.36.0-preview.5 by @gemini-cli-robot in
-  [#24046](https://github.com/google-gemini/gemini-cli/pull/24046)
-- Fix test flakes by globally mocking ink-spinner by @jacob314 in
-  [#24044](https://github.com/google-gemini/gemini-cli/pull/24044)
-- Enable network access in sandbox configuration by @galz10 in
-  [#24055](https://github.com/google-gemini/gemini-cli/pull/24055)
-- feat(context): add configurable memoryBoundaryMarkers setting by @SandyTao520
-  in [#24020](https://github.com/google-gemini/gemini-cli/pull/24020)
-- feat(core): implement windows sandbox expansion and denial detection by
-  @scidomino in [#24027](https://github.com/google-gemini/gemini-cli/pull/24027)
-- fix(core): resolve ACP Operation Aborted Errors in grep_search by @ivanporty
-  in [#23821](https://github.com/google-gemini/gemini-cli/pull/23821)
-- fix(hooks): prevent SessionEnd from firing twice in non-interactive mode by
-  @krishdef7 in [#22139](https://github.com/google-gemini/gemini-cli/pull/22139)
-- Re-word intro to Gemini 3 page. by @g-samroberts in
-  [#24069](https://github.com/google-gemini/gemini-cli/pull/24069)
-- fix(cli): resolve layout contention and flashing loop in StatusRow by
-  @keithguerin in
-  [#24065](https://github.com/google-gemini/gemini-cli/pull/24065)
-- fix(sandbox): implement Windows Mandatory Integrity Control for GeminiSandbox
-  by @galz10 in [#24057](https://github.com/google-gemini/gemini-cli/pull/24057)
-- feat(core): implement tool-based topic grouping (Chapters) by @Abhijit-2592 in
-  [#23150](https://github.com/google-gemini/gemini-cli/pull/23150)
-- feat(cli): support 'tab to queue' for messages while generating by @gundermanc
-  in [#24052](https://github.com/google-gemini/gemini-cli/pull/24052)
-- feat(core): agnostic background task UI with CompletionBehavior by
-  @adamfweidman in
-  [#22740](https://github.com/google-gemini/gemini-cli/pull/22740)
-- UX for topic narration tool by @gundermanc in
-  [#24079](https://github.com/google-gemini/gemini-cli/pull/24079)
-- fix: shellcheck warnings in scripts by @scidomino in
-  [#24035](https://github.com/google-gemini/gemini-cli/pull/24035)
-- test(evals): add comprehensive subagent delegation evaluations by @abhipatel12
-  in [#24132](https://github.com/google-gemini/gemini-cli/pull/24132)
-- fix(a2a-server): prioritize ADC before evaluating headless constraints for
-  auth initialization by @spencer426 in
-  [#23614](https://github.com/google-gemini/gemini-cli/pull/23614)
-- Text can be added after /plan command by @rambleraptor in
-  [#22833](https://github.com/google-gemini/gemini-cli/pull/22833)
-- fix(cli): resolve missing F12 logs via global console store by @scidomino in
-  [#24235](https://github.com/google-gemini/gemini-cli/pull/24235)
-- fix broken tests by @scidomino in
-  [#24279](https://github.com/google-gemini/gemini-cli/pull/24279)
-- fix(evals): add update_topic behavioral eval by @gundermanc in
-  [#24223](https://github.com/google-gemini/gemini-cli/pull/24223)
-- feat(core): Unified Context Management and Tool Distillation. by @joshualitt
-  in [#24157](https://github.com/google-gemini/gemini-cli/pull/24157)
-- Default enable narration for the team. by @gundermanc in
-  [#24224](https://github.com/google-gemini/gemini-cli/pull/24224)
-- fix(core): ensure default agents provide tools and use model-specific schemas
-  by @abhipatel12 in
-  [#24268](https://github.com/google-gemini/gemini-cli/pull/24268)
-- feat(cli): show Flash Lite Preview model regardless of user tier by @sehoon38
-  in [#23904](https://github.com/google-gemini/gemini-cli/pull/23904)
-- feat(cli): implement compact tool output by @jwhelangoog in
-  [#20974](https://github.com/google-gemini/gemini-cli/pull/20974)
-- Add security settings for tool sandboxing by @galz10 in
-  [#23923](https://github.com/google-gemini/gemini-cli/pull/23923)
-- chore(test-utils): switch integration tests to use PREVIEW_GEMINI_MODEL by
-  @sehoon38 in [#24276](https://github.com/google-gemini/gemini-cli/pull/24276)
-- feat(core): enable topic update narration for legacy models by @Abhijit-2592
-  in [#24241](https://github.com/google-gemini/gemini-cli/pull/24241)
-- feat(core): add project-level memory scope to save_memory tool by @SandyTao520
-  in [#24161](https://github.com/google-gemini/gemini-cli/pull/24161)
-- test(integration): fix plan mode write denial test false positive by @sehoon38
-  in [#24299](https://github.com/google-gemini/gemini-cli/pull/24299)
-- feat(plan): support `Plan` mode in untrusted folders by @Adib234 in
-  [#17586](https://github.com/google-gemini/gemini-cli/pull/17586)
-- fix(core): enable mid-stream retries for all models and re-enable compression
-  test by @sehoon38 in
-  [#24302](https://github.com/google-gemini/gemini-cli/pull/24302)
-- Changelog for v0.36.0-preview.6 by @gemini-cli-robot in
-  [#24082](https://github.com/google-gemini/gemini-cli/pull/24082)
-- Changelog for v0.35.3 by @gemini-cli-robot in
-  [#24083](https://github.com/google-gemini/gemini-cli/pull/24083)
-- feat(cli): add auth info to footer by @sehoon38 in
-  [#24042](https://github.com/google-gemini/gemini-cli/pull/24042)
-- fix(browser): reset action counter for each agent session and let it ignore
-  internal actions by @cynthialong0-0 in
-  [#24228](https://github.com/google-gemini/gemini-cli/pull/24228)
-- feat(plan): promote planning feature to stable by @ruomengz in
-  [#24282](https://github.com/google-gemini/gemini-cli/pull/24282)
-- fix(browser): terminate subagent immediately on domain restriction violations
-  by @gsquared94 in
-  [#24313](https://github.com/google-gemini/gemini-cli/pull/24313)
-- feat(cli): add UI to update extensions by @ruomengz in
-  [#23682](https://github.com/google-gemini/gemini-cli/pull/23682)
-- Fix(browser): terminate immediately for "browser is already running" error by
-  @cynthialong0-0 in
-  [#24233](https://github.com/google-gemini/gemini-cli/pull/24233)
-- docs: Add 'plan' option to approval mode in CLI reference by @YifanRuan in
-  [#24134](https://github.com/google-gemini/gemini-cli/pull/24134)
-- fix(core): batch macOS seatbelt rules into a profile file to prevent ARG_MAX
-  errors by @ehedlund in
-  [#24255](https://github.com/google-gemini/gemini-cli/pull/24255)
-- fix(core): fix race condition between browser agent and main closing process
-  by @cynthialong0-0 in
-  [#24340](https://github.com/google-gemini/gemini-cli/pull/24340)
-- perf(build): optimize build scripts for parallel execution and remove
-  redundant checks by @sehoon38 in
-  [#24307](https://github.com/google-gemini/gemini-cli/pull/24307)
-- ci: install bubblewrap on Linux for release workflows by @ehedlund in
-  [#24347](https://github.com/google-gemini/gemini-cli/pull/24347)
-- chore(release): allow bundling for all builds, including stable by @sehoon38
-  in [#24305](https://github.com/google-gemini/gemini-cli/pull/24305)
-- Revert "Add security settings for tool sandboxing" by @jerop in
-  [#24357](https://github.com/google-gemini/gemini-cli/pull/24357)
-- docs: update subagents docs to not be experimental by @abhipatel12 in
-  [#24343](https://github.com/google-gemini/gemini-cli/pull/24343)
-- fix(core): implement **read and **write commands in sandbox managers by
-  @galz10 in [#24283](https://github.com/google-gemini/gemini-cli/pull/24283)
-- don't try to remove tags in dry run by @scidomino in
-  [#24356](https://github.com/google-gemini/gemini-cli/pull/24356)
-- fix(config): disable JIT context loading by default by @SandyTao520 in
-  [#24364](https://github.com/google-gemini/gemini-cli/pull/24364)
-- test(sandbox): add integration test for dynamic permission expansion by
-  @galz10 in [#24359](https://github.com/google-gemini/gemini-cli/pull/24359)
-- docs(policy): remove unsupported mcpName wildcard edge case by @abhipatel12 in
-  [#24133](https://github.com/google-gemini/gemini-cli/pull/24133)
-- docs: fix broken GEMINI.md link in CONTRIBUTING.md by @Panchal-Tirth in
-  [#24182](https://github.com/google-gemini/gemini-cli/pull/24182)
-- feat(core): infrastructure for event-driven subagent history by @abhipatel12
-  in [#23914](https://github.com/google-gemini/gemini-cli/pull/23914)
-- fix(core): resolve Plan Mode deadlock during plan file creation due to sandbox
-  restrictions by @DavidAPierce in
-  [#24047](https://github.com/google-gemini/gemini-cli/pull/24047)
-- fix(core): fix browser agent UX issues and improve E2E test reliability by
-  @gsquared94 in
-  [#24312](https://github.com/google-gemini/gemini-cli/pull/24312)
-- fix(ui): wrap topic and intent fields in TopicMessage by @jwhelangoog in
-  [#24386](https://github.com/google-gemini/gemini-cli/pull/24386)
-- refactor(core): Centralize context management logic into src/context by
-  @joshualitt in
-  [#24380](https://github.com/google-gemini/gemini-cli/pull/24380)
-- fix(core): pin AuthType.GATEWAY to use Gemini 3.1 Pro/Flash Lite by default by
-  @sripasg in [#24375](https://github.com/google-gemini/gemini-cli/pull/24375)
-- feat(ui): add Tokyo Night theme by @danrneal in
-  [#24054](https://github.com/google-gemini/gemini-cli/pull/24054)
-- fix(cli): refactor test config loading and mock debugLogger in test-setup by
-  @mattKorwel in
-  [#24389](https://github.com/google-gemini/gemini-cli/pull/24389)
-- Set memoryManager to false in settings.json by @mattKorwel in
-  [#24393](https://github.com/google-gemini/gemini-cli/pull/24393)
-- ink 6.6.3 by @jacob314 in
-  [#24372](https://github.com/google-gemini/gemini-cli/pull/24372)
-- fix(core): resolve subagent chat recording gaps and directory inheritance by
+- fix(patch): cherry-pick 050c303 to release/v0.38.0-pr-25317 to patch version
+  v0.38.0 and create version 0.38.1 by @gemini-cli-robot in
+  [#25466](https://github.com/google-gemini/gemini-cli/pull/25466)
+- fix(cli): refresh slash command list after /skills reload by @NTaylorMullen in
+  [#24454](https://github.com/google-gemini/gemini-cli/pull/24454)
+- Update README.md for links. by @g-samroberts in
+  [#22759](https://github.com/google-gemini/gemini-cli/pull/22759)
+- fix(core): ensure complete_task tool calls are recorded in chat history by
   @abhipatel12 in
-  [#24368](https://github.com/google-gemini/gemini-cli/pull/24368)
-- fix(cli): cap shell output at 10 MB to prevent RangeError crash by @ProthamD
-  in [#24168](https://github.com/google-gemini/gemini-cli/pull/24168)
-- feat(plan): conditionally add enter/exit plan mode tools based on current mode
-  by @ruomengz in
-  [#24378](https://github.com/google-gemini/gemini-cli/pull/24378)
-- feat(core): prioritize discussion before formal plan approval by @jerop in
-  [#24423](https://github.com/google-gemini/gemini-cli/pull/24423)
-- fix(ui): add accelerated scrolling on alternate buffer mode by @devr0306 in
-  [#23940](https://github.com/google-gemini/gemini-cli/pull/23940)
-- feat(core): populate sandbox forbidden paths with project ignore file contents
-  by @ehedlund in
-  [#24038](https://github.com/google-gemini/gemini-cli/pull/24038)
-- fix(core): ensure blue border overlay and input blocker to act correctly
-  depending on browser agent activities by @cynthialong0-0 in
-  [#24385](https://github.com/google-gemini/gemini-cli/pull/24385)
-- fix(ui): removed additional vertical padding for tables by @devr0306 in
-  [#24381](https://github.com/google-gemini/gemini-cli/pull/24381)
-- fix(build): upload full bundle directory archive to GitHub releases by
-  @sehoon38 in [#24403](https://github.com/google-gemini/gemini-cli/pull/24403)
-- fix(build): wire bundle:browser-mcp into bundle pipeline by @gsquared94 in
-  [#24424](https://github.com/google-gemini/gemini-cli/pull/24424)
-- feat(browser): add sandbox-aware browser agent initialization by @gsquared94
-  in [#24419](https://github.com/google-gemini/gemini-cli/pull/24419)
-- feat(core): enhance tracker task schemas for detailed titles and descriptions
-  by @anj-s in [#23902](https://github.com/google-gemini/gemini-cli/pull/23902)
-- refactor(core): Unified context management settings schema by @joshualitt in
-  [#24391](https://github.com/google-gemini/gemini-cli/pull/24391)
-- feat(core): update browser agent prompt to check open pages first when
-  bringing up by @cynthialong0-0 in
-  [#24431](https://github.com/google-gemini/gemini-cli/pull/24431)
-- fix(acp) refactor(core,cli): centralize model discovery logic in
-  ModelConfigService by @sripasg in
-  [#24392](https://github.com/google-gemini/gemini-cli/pull/24392)
-- Changelog for v0.36.0-preview.7 by @gemini-cli-robot in
-  [#24346](https://github.com/google-gemini/gemini-cli/pull/24346)
-- fix: update task tracker storage location in system prompt by @anj-s in
-  [#24034](https://github.com/google-gemini/gemini-cli/pull/24034)
-- feat(browser): supersede stale snapshots to reclaim context-window tokens by
+  [#24437](https://github.com/google-gemini/gemini-cli/pull/24437)
+- feat(policy): explicitly allow web_fetch in plan mode with ask_user by
+  @Adib234 in [#24456](https://github.com/google-gemini/gemini-cli/pull/24456)
+- fix(core): refactor linux sandbox to fix ARG_MAX crashes by @ehedlund in
+  [#24286](https://github.com/google-gemini/gemini-cli/pull/24286)
+- feat(config): add experimental.adk.agentSessionNoninteractiveEnabled setting
+  by @adamfweidman in
+  [#24439](https://github.com/google-gemini/gemini-cli/pull/24439)
+- Changelog for v0.36.0-preview.8 by @gemini-cli-robot in
+  [#24453](https://github.com/google-gemini/gemini-cli/pull/24453)
+- feat(cli): change default loadingPhrases to 'off' to hide tips by @keithguerin
+  in [#24342](https://github.com/google-gemini/gemini-cli/pull/24342)
+- fix(cli): ensure agent stops when all declinable tools are cancelled by
+  @NTaylorMullen in
+  [#24479](https://github.com/google-gemini/gemini-cli/pull/24479)
+- fix(core): enhance sandbox usability and fix build error by @galz10 in
+  [#24460](https://github.com/google-gemini/gemini-cli/pull/24460)
+- Terminal Serializer Optimization by @jacob314 in
+  [#24485](https://github.com/google-gemini/gemini-cli/pull/24485)
+- Auto configure memory. by @jacob314 in
+  [#24474](https://github.com/google-gemini/gemini-cli/pull/24474)
+- Unused error variables in catch block are not allowed by @alisa-alisa in
+  [#24487](https://github.com/google-gemini/gemini-cli/pull/24487)
+- feat(core): add background memory service for skill extraction by @SandyTao520
+  in [#24274](https://github.com/google-gemini/gemini-cli/pull/24274)
+- feat: implement high-signal PR regression check for evaluations by
+  @alisa-alisa in
+  [#23937](https://github.com/google-gemini/gemini-cli/pull/23937)
+- Fix shell output display by @jacob314 in
+  [#24490](https://github.com/google-gemini/gemini-cli/pull/24490)
+- fix(ui): resolve unwanted vertical spacing around various tool output
+  treatments by @jwhelangoog in
+  [#24449](https://github.com/google-gemini/gemini-cli/pull/24449)
+- revert(cli): bring back input box and footer visibility in copy mode by
+  @sehoon38 in [#24504](https://github.com/google-gemini/gemini-cli/pull/24504)
+- fix(cli): prevent crash in AnsiOutputText when handling non-array data by
+  @sehoon38 in [#24498](https://github.com/google-gemini/gemini-cli/pull/24498)
+- feat(cli): support default values for environment variables by @ruomengz in
+  [#24469](https://github.com/google-gemini/gemini-cli/pull/24469)
+- Implement background process monitoring and inspection tools by @cocosheng-g
+  in [#23799](https://github.com/google-gemini/gemini-cli/pull/23799)
+- docs(browser-agent): update stale browser agent documentation by @gsquared94
+  in [#24463](https://github.com/google-gemini/gemini-cli/pull/24463)
+- fix: enable browser_agent in integration tests and add localhost fixture tests
+  by @gsquared94 in
+  [#24523](https://github.com/google-gemini/gemini-cli/pull/24523)
+- fix(browser): handle computer-use model detection for analyze_screenshot by
   @gsquared94 in
-  [#24440](https://github.com/google-gemini/gemini-cli/pull/24440)
-- docs(core): add subagent tool isolation draft doc by @akh64bit in
-  [#23275](https://github.com/google-gemini/gemini-cli/pull/23275)
-- fix(patch): cherry-pick 64c928f to release/v0.37.0-preview.0-pr-23257 to patch
-  version v0.37.0-preview.0 and create version 0.37.0-preview.1 by
-  @gemini-cli-robot in
-  [#24561](https://github.com/google-gemini/gemini-cli/pull/24561)
-- fix(patch): cherry-pick cb7f7d6 to release/v0.37.0-preview.1-pr-24342 to patch
-  version v0.37.0-preview.1 and create version 0.37.0-preview.2 by
-  @gemini-cli-robot in
-  [#24842](https://github.com/google-gemini/gemini-cli/pull/24842)
+  [#24502](https://github.com/google-gemini/gemini-cli/pull/24502)
+- feat(core): Land ContextCompressionService by @joshualitt in
+  [#24483](https://github.com/google-gemini/gemini-cli/pull/24483)
+- feat(core): scope subagent workspace directories via AsyncLocalStorage by
+  @SandyTao520 in
+  [#24445](https://github.com/google-gemini/gemini-cli/pull/24445)
+- Update ink version to 6.6.7 by @jacob314 in
+  [#24514](https://github.com/google-gemini/gemini-cli/pull/24514)
+- fix(acp): handle all InvalidStreamError types gracefully in prompt by @sripasg
+  in [#24540](https://github.com/google-gemini/gemini-cli/pull/24540)
+- Fix crash when vim editor is not found in PATH on Windows by
+  @Nagajyothi-tammisetti in
+  [#22423](https://github.com/google-gemini/gemini-cli/pull/22423)
+- fix(core): move project memory dir under tmp directory by @SandyTao520 in
+  [#24542](https://github.com/google-gemini/gemini-cli/pull/24542)
+- Enable 'Other' option for yesno question type by @ruomengz in
+  [#24545](https://github.com/google-gemini/gemini-cli/pull/24545)
+- fix(cli): clear stale retry/loading state after cancellation (#21096) by
+  @Aaxhirrr in [#21960](https://github.com/google-gemini/gemini-cli/pull/21960)
+- Changelog for v0.37.0-preview.0 by @gemini-cli-robot in
+  [#24464](https://github.com/google-gemini/gemini-cli/pull/24464)
+- feat(core): implement context-aware persistent policy approvals by @jerop in
+  [#23257](https://github.com/google-gemini/gemini-cli/pull/23257)
+- docs: move agent disabling instructions and update remote agent status by
+  @jackwotherspoon in
+  [#24559](https://github.com/google-gemini/gemini-cli/pull/24559)
+- feat(cli): migrate nonInteractiveCli to LegacyAgentSession by @adamfweidman in
+  [#22987](https://github.com/google-gemini/gemini-cli/pull/22987)
+- fix(core): unsafe type assertions in Core File System #19712 by
+  @aniketsaurav18 in
+  [#19739](https://github.com/google-gemini/gemini-cli/pull/19739)
+- fix(ui): hide model quota in /stats and refactor quota display by @danzaharia1
+  in [#24206](https://github.com/google-gemini/gemini-cli/pull/24206)
+- Changelog for v0.36.0 by @gemini-cli-robot in
+  [#24558](https://github.com/google-gemini/gemini-cli/pull/24558)
+- Changelog for v0.37.0-preview.1 by @gemini-cli-robot in
+  [#24568](https://github.com/google-gemini/gemini-cli/pull/24568)
+- docs: add missing .md extensions to internal doc links by @ishaan-arora-1 in
+  [#24145](https://github.com/google-gemini/gemini-cli/pull/24145)
+- fix(ui): fixed table styling by @devr0306 in
+  [#24565](https://github.com/google-gemini/gemini-cli/pull/24565)
+- fix(core): pass includeDirectories to sandbox configuration by @galz10 in
+  [#24573](https://github.com/google-gemini/gemini-cli/pull/24573)
+- feat(ui): enable "TerminalBuffer" mode to solve flicker by @jacob314 in
+  [#24512](https://github.com/google-gemini/gemini-cli/pull/24512)
+- docs: clarify release coordination by @scidomino in
+  [#24575](https://github.com/google-gemini/gemini-cli/pull/24575)
+- fix(core): remove broken PowerShell translation and fix native \_\_write in
+  Windows sandbox by @scidomino in
+  [#24571](https://github.com/google-gemini/gemini-cli/pull/24571)
+- Add instructions for how to start react in prod and force react to prod mode
+  by @jacob314 in
+  [#24590](https://github.com/google-gemini/gemini-cli/pull/24590)
+- feat(cli): minimalist sandbox status labels by @galz10 in
+  [#24582](https://github.com/google-gemini/gemini-cli/pull/24582)
+- Feat/browser agent metrics by @kunal-10-cloud in
+  [#24210](https://github.com/google-gemini/gemini-cli/pull/24210)
+- test: fix Windows CI execution and resolve exposed platform failures by
+  @ehedlund in [#24476](https://github.com/google-gemini/gemini-cli/pull/24476)
+- feat(core,cli): prioritize summary for topics (#24608) by @Abhijit-2592 in
+  [#24609](https://github.com/google-gemini/gemini-cli/pull/24609)
+- show color by @jacob314 in
+  [#24613](https://github.com/google-gemini/gemini-cli/pull/24613)
+- feat(cli): enable compact tool output by default (#24509) by @jwhelangoog in
+  [#24510](https://github.com/google-gemini/gemini-cli/pull/24510)
+- fix(core): inject skill system instructions into subagent prompts if activated
+  by @abhipatel12 in
+  [#24620](https://github.com/google-gemini/gemini-cli/pull/24620)
+- fix(core): improve windows sandbox reliability and fix integration tests by
+  @ehedlund in [#24480](https://github.com/google-gemini/gemini-cli/pull/24480)
+- fix(core): ensure sandbox approvals are correctly persisted and matched for
+  proactive expansions by @galz10 in
+  [#24577](https://github.com/google-gemini/gemini-cli/pull/24577)
+- feat(cli) Scrollbar for input prompt by @jacob314 in
+  [#21992](https://github.com/google-gemini/gemini-cli/pull/21992)
+- Do not run pr-eval workflow when no steering changes detected by @alisa-alisa
+  in [#24621](https://github.com/google-gemini/gemini-cli/pull/24621)
+- Fix restoration of topic headers. by @gundermanc in
+  [#24650](https://github.com/google-gemini/gemini-cli/pull/24650)
+- feat(core): discourage update topic tool for simple tasks by @Samee24 in
+  [#24640](https://github.com/google-gemini/gemini-cli/pull/24640)
+- fix(core): ensure global temp directory is always in sandbox allowed paths by
+  @galz10 in [#24638](https://github.com/google-gemini/gemini-cli/pull/24638)
+- fix(core): detect uninitialized lines by @jacob314 in
+  [#24646](https://github.com/google-gemini/gemini-cli/pull/24646)
+- docs: update sandboxing documentation and toolSandboxing settings by @galz10
+  in [#24655](https://github.com/google-gemini/gemini-cli/pull/24655)
+- feat(cli): enhance tool confirmation UI and selection layout by @galz10 in
+  [#24376](https://github.com/google-gemini/gemini-cli/pull/24376)
+- feat(acp): add support for `/about` command by @sripasg in
+  [#24649](https://github.com/google-gemini/gemini-cli/pull/24649)
+- feat(cli): add role specific metrics to /stats by @cynthialong0-0 in
+  [#24659](https://github.com/google-gemini/gemini-cli/pull/24659)
+- split context by @jacob314 in
+  [#24623](https://github.com/google-gemini/gemini-cli/pull/24623)
+- fix(cli): remove -S from shebang to fix Windows and BSD execution by
+  @scidomino in [#24756](https://github.com/google-gemini/gemini-cli/pull/24756)
+- Fix issue where topic headers can be posted back to back by @gundermanc in
+  [#24759](https://github.com/google-gemini/gemini-cli/pull/24759)
+- fix(core): handle partial llm_request in BeforeModel hook override by
+  @krishdef7 in [#22326](https://github.com/google-gemini/gemini-cli/pull/22326)
+- fix(ui): improve narration suppression and reduce flicker by @gundermanc in
+  [#24635](https://github.com/google-gemini/gemini-cli/pull/24635)
+- fix(ui): fixed auth race condition causing logo to flicker by @devr0306 in
+  [#24652](https://github.com/google-gemini/gemini-cli/pull/24652)
+- fix(browser): remove premature browser cleanup after subagent invocation by
+  @gsquared94 in
+  [#24753](https://github.com/google-gemini/gemini-cli/pull/24753)
+- Revert "feat(core,cli): prioritize summary for topics (#24608)" by
+  @Abhijit-2592 in
+  [#24777](https://github.com/google-gemini/gemini-cli/pull/24777)
+- relax tool sandboxing overrides for plan mode to match defaults. by
+  @DavidAPierce in
+  [#24762](https://github.com/google-gemini/gemini-cli/pull/24762)
+- fix(cli): respect global environment variable allowlist by @scidomino in
+  [#24767](https://github.com/google-gemini/gemini-cli/pull/24767)
+- fix(cli): ensure skills list outputs to stdout in non-interactive environments
+  by @spencer426 in
+  [#24566](https://github.com/google-gemini/gemini-cli/pull/24566)
+- Add an eval for and fix unsafe cloning behavior. by @gundermanc in
+  [#24457](https://github.com/google-gemini/gemini-cli/pull/24457)
+- fix(policy): allow complete_task in plan mode by @abhipatel12 in
+  [#24771](https://github.com/google-gemini/gemini-cli/pull/24771)
+- feat(telemetry): add browser agent clearcut metrics by @gsquared94 in
+  [#24688](https://github.com/google-gemini/gemini-cli/pull/24688)
+- feat(cli): support selective topic expansion and click-to-expand by
+  @Abhijit-2592 in
+  [#24793](https://github.com/google-gemini/gemini-cli/pull/24793)
+- temporarily disable sandbox integration test on windows by @ehedlund in
+  [#24786](https://github.com/google-gemini/gemini-cli/pull/24786)
+- Remove flakey test by @scidomino in
+  [#24837](https://github.com/google-gemini/gemini-cli/pull/24837)
+- Alisa/approve button by @alisa-alisa in
+  [#24645](https://github.com/google-gemini/gemini-cli/pull/24645)
+- feat(hooks): display hook system messages in UI by @mbleigh in
+  [#24616](https://github.com/google-gemini/gemini-cli/pull/24616)
+- fix(core): propagate BeforeModel hook model override end-to-end by @krishdef7
+  in [#24784](https://github.com/google-gemini/gemini-cli/pull/24784)
+- chore: fix formatting for behavioral eval skill reference file by @abhipatel12
+  in [#24846](https://github.com/google-gemini/gemini-cli/pull/24846)
+- fix: use directory junctions on Windows for skill linking by @enjoykumawat in
+  [#24823](https://github.com/google-gemini/gemini-cli/pull/24823)
+- fix(cli): prevent multiple banner increments on remount by @sehoon38 in
+  [#24843](https://github.com/google-gemini/gemini-cli/pull/24843)
+- feat(acp): add /help command by @sripasg in
+  [#24839](https://github.com/google-gemini/gemini-cli/pull/24839)
+- fix(core): remove tmux alternate buffer warning by @jackwotherspoon in
+  [#24852](https://github.com/google-gemini/gemini-cli/pull/24852)
+- Improve sandbox error matching and caching by @DavidAPierce in
+  [#24550](https://github.com/google-gemini/gemini-cli/pull/24550)
+- feat(core): add agent protocol UI types and experimental flag by @mbleigh in
+  [#24275](https://github.com/google-gemini/gemini-cli/pull/24275)
+- feat(core): use experiment flags for default fetch timeouts by @yunaseoul in
+  [#24261](https://github.com/google-gemini/gemini-cli/pull/24261)
+- Revert "fix(ui): improve narration suppression and reduce flicker (#2… by
+  @gundermanc in
+  [#24857](https://github.com/google-gemini/gemini-cli/pull/24857)
+- refactor(cli): remove duplication in interactive shell awaiting input hint by
+  @JayadityaGit in
+  [#24801](https://github.com/google-gemini/gemini-cli/pull/24801)
+- refactor(core): make LegacyAgentSession dependencies optional by @mbleigh in
+  [#24287](https://github.com/google-gemini/gemini-cli/pull/24287)
+- Changelog for v0.37.0-preview.2 by @gemini-cli-robot in
+  [#24848](https://github.com/google-gemini/gemini-cli/pull/24848)
+- fix(cli): always show shell command description or actual command by @jacob314
+  in [#24774](https://github.com/google-gemini/gemini-cli/pull/24774)
+- Added flag for ept size and increased default size by @devr0306 in
+  [#24859](https://github.com/google-gemini/gemini-cli/pull/24859)
+- fix(core): dispose Scheduler to prevent McpProgress listener leak by
+  @Anjaligarhwal in
+  [#24870](https://github.com/google-gemini/gemini-cli/pull/24870)
+- fix(cli): switch default back to terminalBuffer=false and fix regressions
+  introduced for that mode by @jacob314 in
+  [#24873](https://github.com/google-gemini/gemini-cli/pull/24873)
+- feat(cli): switch to ctrl+g from ctrl-x by @jacob314 in
+  [#24861](https://github.com/google-gemini/gemini-cli/pull/24861)
+- fix: isolate concurrent browser agent instances by @gsquared94 in
+  [#24794](https://github.com/google-gemini/gemini-cli/pull/24794)
+- docs: update MCP server OAuth redirect port documentation by @adamfweidman in
+  [#24844](https://github.com/google-gemini/gemini-cli/pull/24844)
 
 **Full Changelog**:
-https://github.com/google-gemini/gemini-cli/compare/v0.36.0...v0.37.1
+https://github.com/google-gemini/gemini-cli/compare/v0.38.0...v0.38.1

docs/changelogs/preview.md

@@ -1,6 +1,6 @@
-# Preview release: v0.38.0-preview.0
+# Preview release: v0.39.0-preview.0
 
-Released: April 08, 2026
+Released: April 14, 2026
 
 Our preview release includes the latest, new, and experimental features. This
 release may not be as stable as our [latest weekly release](latest.md).
@@ -13,256 +13,245 @@ npm install -g @google/gemini-cli@preview
 
 ## Highlights
 
-- **Context Management:** Introduced a Context Compression Service to optimize
-  context window usage and landed a background memory service for skill
-  extraction.
-- **Enhanced Security:** Implemented context-aware persistent policy approvals
-  for smarter tool permissions and enabled `web_fetch` in plan mode with user
-  confirmation.
-- **Workflow Monitoring:** Added background process monitoring and inspection
-  tools for better visibility into long-running tasks.
-- **UI/UX Refinements:** Enhanced the tool confirmation UI, selection layout,
-  and added support for selective topic expansion and click-to-expand.
-- **Core Stability:** Improved sandbox reliability on Linux and Windows,
-  resolved shebang compatibility issues, and fixed various crashes in the CLI
-  and core services.
+- **Refactored Subagents and Unified Tooling:** Consolidate subagent tools into
+  a single `invoke_subagent` tool, removed legacy wrapping tools, and improved
+  turn limits for codebase investigator.
+- **Advanced Memory and Skill Management:** Introduced `/memory` inbox for
+  reviewing extracted skills and added skill patching support, enhancing agent
+  learning and persistence.
+- **Expanded Test and Evaluation Infrastructure:** Added memory and CPU
+  performance integration test harnesses and generalized evaluation
+  infrastructure for better suite organization.
+- **Sandbox and Security Hardening:** Centralized sandbox paths for Linux and
+  macOS, enforced read-only security for async git worktree resolution, and
+  optimized Windows sandbox initialization.
+- **Enhanced CLI UX and UI Stability:** Improved scroll momentum, added a
+  `debugRainbow` setting, and resolved various memory leaks and PTY exhaustion
+  issues for a smoother terminal experience.
 
 ## What's Changed
 
-- fix(cli): refresh slash command list after /skills reload by @NTaylorMullen in
-  [#24454](https://github.com/google-gemini/gemini-cli/pull/24454)
-- Update README.md for links. by @g-samroberts in
-  [#22759](https://github.com/google-gemini/gemini-cli/pull/22759)
-- fix(core): ensure complete_task tool calls are recorded in chat history by
-  @abhipatel12 in
-  [#24437](https://github.com/google-gemini/gemini-cli/pull/24437)
-- feat(policy): explicitly allow web_fetch in plan mode with ask_user by
-  @Adib234 in [#24456](https://github.com/google-gemini/gemini-cli/pull/24456)
-- fix(core): refactor linux sandbox to fix ARG_MAX crashes by @ehedlund in
-  [#24286](https://github.com/google-gemini/gemini-cli/pull/24286)
-- feat(config): add experimental.adk.agentSessionNoninteractiveEnabled setting
-  by @adamfweidman in
-  [#24439](https://github.com/google-gemini/gemini-cli/pull/24439)
-- Changelog for v0.36.0-preview.8 by @gemini-cli-robot in
-  [#24453](https://github.com/google-gemini/gemini-cli/pull/24453)
-- feat(cli): change default loadingPhrases to 'off' to hide tips by @keithguerin
-  in [#24342](https://github.com/google-gemini/gemini-cli/pull/24342)
-- fix(cli): ensure agent stops when all declinable tools are cancelled by
-  @NTaylorMullen in
-  [#24479](https://github.com/google-gemini/gemini-cli/pull/24479)
-- fix(core): enhance sandbox usability and fix build error by @galz10 in
-  [#24460](https://github.com/google-gemini/gemini-cli/pull/24460)
-- Terminal Serializer Optimization by @jacob314 in
-  [#24485](https://github.com/google-gemini/gemini-cli/pull/24485)
-- Auto configure memory. by @jacob314 in
-  [#24474](https://github.com/google-gemini/gemini-cli/pull/24474)
-- Unused error variables in catch block are not allowed by @alisa-alisa in
-  [#24487](https://github.com/google-gemini/gemini-cli/pull/24487)
-- feat(core): add background memory service for skill extraction by @SandyTao520
-  in [#24274](https://github.com/google-gemini/gemini-cli/pull/24274)
-- feat: implement high-signal PR regression check for evaluations by
-  @alisa-alisa in
-  [#23937](https://github.com/google-gemini/gemini-cli/pull/23937)
-- Fix shell output display by @jacob314 in
-  [#24490](https://github.com/google-gemini/gemini-cli/pull/24490)
-- fix(ui): resolve unwanted vertical spacing around various tool output
-  treatments by @jwhelangoog in
-  [#24449](https://github.com/google-gemini/gemini-cli/pull/24449)
-- revert(cli): bring back input box and footer visibility in copy mode by
-  @sehoon38 in [#24504](https://github.com/google-gemini/gemini-cli/pull/24504)
-- fix(cli): prevent crash in AnsiOutputText when handling non-array data by
-  @sehoon38 in [#24498](https://github.com/google-gemini/gemini-cli/pull/24498)
-- feat(cli): support default values for environment variables by @ruomengz in
-  [#24469](https://github.com/google-gemini/gemini-cli/pull/24469)
-- Implement background process monitoring and inspection tools by @cocosheng-g
-  in [#23799](https://github.com/google-gemini/gemini-cli/pull/23799)
-- docs(browser-agent): update stale browser agent documentation by @gsquared94
-  in [#24463](https://github.com/google-gemini/gemini-cli/pull/24463)
-- fix: enable browser_agent in integration tests and add localhost fixture tests
-  by @gsquared94 in
-  [#24523](https://github.com/google-gemini/gemini-cli/pull/24523)
-- fix(browser): handle computer-use model detection for analyze_screenshot by
-  @gsquared94 in
-  [#24502](https://github.com/google-gemini/gemini-cli/pull/24502)
-- feat(core): Land ContextCompressionService by @joshualitt in
-  [#24483](https://github.com/google-gemini/gemini-cli/pull/24483)
-- feat(core): scope subagent workspace directories via AsyncLocalStorage by
+- refactor(plan): simplify policy priorities and consolidate read-only rules by
+  @ruomengz in [#24849](https://github.com/google-gemini/gemini-cli/pull/24849)
+- feat(test-utils): add memory usage integration test harness by @sripasg in
+  [#24876](https://github.com/google-gemini/gemini-cli/pull/24876)
+- feat(memory): add /memory inbox command for reviewing extracted skills by
   @SandyTao520 in
-  [#24445](https://github.com/google-gemini/gemini-cli/pull/24445)
-- Update ink version to 6.6.7 by @jacob314 in
-  [#24514](https://github.com/google-gemini/gemini-cli/pull/24514)
-- fix(acp): handle all InvalidStreamError types gracefully in prompt by @sripasg
-  in [#24540](https://github.com/google-gemini/gemini-cli/pull/24540)
-- Fix crash when vim editor is not found in PATH on Windows by
-  @Nagajyothi-tammisetti in
-  [#22423](https://github.com/google-gemini/gemini-cli/pull/22423)
-- fix(core): move project memory dir under tmp directory by @SandyTao520 in
-  [#24542](https://github.com/google-gemini/gemini-cli/pull/24542)
-- Enable 'Other' option for yesno question type by @ruomengz in
-  [#24545](https://github.com/google-gemini/gemini-cli/pull/24545)
-- fix(cli): clear stale retry/loading state after cancellation (#21096) by
-  @Aaxhirrr in [#21960](https://github.com/google-gemini/gemini-cli/pull/21960)
-- Changelog for v0.37.0-preview.0 by @gemini-cli-robot in
-  [#24464](https://github.com/google-gemini/gemini-cli/pull/24464)
-- feat(core): implement context-aware persistent policy approvals by @jerop in
-  [#23257](https://github.com/google-gemini/gemini-cli/pull/23257)
-- docs: move agent disabling instructions and update remote agent status by
-  @jackwotherspoon in
-  [#24559](https://github.com/google-gemini/gemini-cli/pull/24559)
-- feat(cli): migrate nonInteractiveCli to LegacyAgentSession by @adamfweidman in
-  [#22987](https://github.com/google-gemini/gemini-cli/pull/22987)
-- fix(core): unsafe type assertions in Core File System #19712 by
-  @aniketsaurav18 in
-  [#19739](https://github.com/google-gemini/gemini-cli/pull/19739)
-- fix(ui): hide model quota in /stats and refactor quota display by @danzaharia1
-  in [#24206](https://github.com/google-gemini/gemini-cli/pull/24206)
-- Changelog for v0.36.0 by @gemini-cli-robot in
-  [#24558](https://github.com/google-gemini/gemini-cli/pull/24558)
-- Changelog for v0.37.0-preview.1 by @gemini-cli-robot in
-  [#24568](https://github.com/google-gemini/gemini-cli/pull/24568)
-- docs: add missing .md extensions to internal doc links by @ishaan-arora-1 in
-  [#24145](https://github.com/google-gemini/gemini-cli/pull/24145)
-- fix(ui): fixed table styling by @devr0306 in
-  [#24565](https://github.com/google-gemini/gemini-cli/pull/24565)
-- fix(core): pass includeDirectories to sandbox configuration by @galz10 in
-  [#24573](https://github.com/google-gemini/gemini-cli/pull/24573)
-- feat(ui): enable "TerminalBuffer" mode to solve flicker by @jacob314 in
-  [#24512](https://github.com/google-gemini/gemini-cli/pull/24512)
-- docs: clarify release coordination by @scidomino in
-  [#24575](https://github.com/google-gemini/gemini-cli/pull/24575)
-- fix(core): remove broken PowerShell translation and fix native \_\_write in
-  Windows sandbox by @scidomino in
-  [#24571](https://github.com/google-gemini/gemini-cli/pull/24571)
-- Add instructions for how to start react in prod and force react to prod mode
-  by @jacob314 in
-  [#24590](https://github.com/google-gemini/gemini-cli/pull/24590)
-- feat(cli): minimalist sandbox status labels by @galz10 in
-  [#24582](https://github.com/google-gemini/gemini-cli/pull/24582)
-- Feat/browser agent metrics by @kunal-10-cloud in
-  [#24210](https://github.com/google-gemini/gemini-cli/pull/24210)
-- test: fix Windows CI execution and resolve exposed platform failures by
-  @ehedlund in [#24476](https://github.com/google-gemini/gemini-cli/pull/24476)
-- feat(core,cli): prioritize summary for topics (#24608) by @Abhijit-2592 in
-  [#24609](https://github.com/google-gemini/gemini-cli/pull/24609)
-- show color by @jacob314 in
-  [#24613](https://github.com/google-gemini/gemini-cli/pull/24613)
-- feat(cli): enable compact tool output by default (#24509) by @jwhelangoog in
-  [#24510](https://github.com/google-gemini/gemini-cli/pull/24510)
-- fix(core): inject skill system instructions into subagent prompts if activated
-  by @abhipatel12 in
-  [#24620](https://github.com/google-gemini/gemini-cli/pull/24620)
-- fix(core): improve windows sandbox reliability and fix integration tests by
-  @ehedlund in [#24480](https://github.com/google-gemini/gemini-cli/pull/24480)
-- fix(core): ensure sandbox approvals are correctly persisted and matched for
-  proactive expansions by @galz10 in
-  [#24577](https://github.com/google-gemini/gemini-cli/pull/24577)
-- feat(cli) Scrollbar for input prompt by @jacob314 in
-  [#21992](https://github.com/google-gemini/gemini-cli/pull/21992)
-- Do not run pr-eval workflow when no steering changes detected by @alisa-alisa
-  in [#24621](https://github.com/google-gemini/gemini-cli/pull/24621)
-- Fix restoration of topic headers. by @gundermanc in
-  [#24650](https://github.com/google-gemini/gemini-cli/pull/24650)
-- feat(core): discourage update topic tool for simple tasks by @Samee24 in
-  [#24640](https://github.com/google-gemini/gemini-cli/pull/24640)
-- fix(core): ensure global temp directory is always in sandbox allowed paths by
-  @galz10 in [#24638](https://github.com/google-gemini/gemini-cli/pull/24638)
-- fix(core): detect uninitialized lines by @jacob314 in
-  [#24646](https://github.com/google-gemini/gemini-cli/pull/24646)
-- docs: update sandboxing documentation and toolSandboxing settings by @galz10
-  in [#24655](https://github.com/google-gemini/gemini-cli/pull/24655)
-- feat(cli): enhance tool confirmation UI and selection layout by @galz10 in
-  [#24376](https://github.com/google-gemini/gemini-cli/pull/24376)
-- feat(acp): add support for `/about` command by @sripasg in
-  [#24649](https://github.com/google-gemini/gemini-cli/pull/24649)
-- feat(cli): add role specific metrics to /stats by @cynthialong0-0 in
-  [#24659](https://github.com/google-gemini/gemini-cli/pull/24659)
-- split context by @jacob314 in
-  [#24623](https://github.com/google-gemini/gemini-cli/pull/24623)
-- fix(cli): remove -S from shebang to fix Windows and BSD execution by
-  @scidomino in [#24756](https://github.com/google-gemini/gemini-cli/pull/24756)
-- Fix issue where topic headers can be posted back to back by @gundermanc in
-  [#24759](https://github.com/google-gemini/gemini-cli/pull/24759)
-- fix(core): handle partial llm_request in BeforeModel hook override by
-  @krishdef7 in [#22326](https://github.com/google-gemini/gemini-cli/pull/22326)
-- fix(ui): improve narration suppression and reduce flicker by @gundermanc in
-  [#24635](https://github.com/google-gemini/gemini-cli/pull/24635)
-- fix(ui): fixed auth race condition causing logo to flicker by @devr0306 in
-  [#24652](https://github.com/google-gemini/gemini-cli/pull/24652)
-- fix(browser): remove premature browser cleanup after subagent invocation by
-  @gsquared94 in
-  [#24753](https://github.com/google-gemini/gemini-cli/pull/24753)
-- Revert "feat(core,cli): prioritize summary for topics (#24608)" by
-  @Abhijit-2592 in
-  [#24777](https://github.com/google-gemini/gemini-cli/pull/24777)
-- relax tool sandboxing overrides for plan mode to match defaults. by
-  @DavidAPierce in
-  [#24762](https://github.com/google-gemini/gemini-cli/pull/24762)
-- fix(cli): respect global environment variable allowlist by @scidomino in
-  [#24767](https://github.com/google-gemini/gemini-cli/pull/24767)
-- fix(cli): ensure skills list outputs to stdout in non-interactive environments
+  [#24544](https://github.com/google-gemini/gemini-cli/pull/24544)
+- chore(release): bump version to 0.39.0-nightly.20260408.e77b22e63 by
+  @gemini-cli-robot in
+  [#24939](https://github.com/google-gemini/gemini-cli/pull/24939)
+- fix(core): ensure robust sandbox cleanup in all process execution paths by
+  @ehedlund in [#24763](https://github.com/google-gemini/gemini-cli/pull/24763)
+- chore: update ink version to 6.6.8 by @jacob314 in
+  [#24934](https://github.com/google-gemini/gemini-cli/pull/24934)
+- Changelog for v0.38.0-preview.0 by @gemini-cli-robot in
+  [#24938](https://github.com/google-gemini/gemini-cli/pull/24938)
+- chore: ignore conductor directory by @JayadityaGit in
+  [#22128](https://github.com/google-gemini/gemini-cli/pull/22128)
+- Changelog for v0.37.0 by @gemini-cli-robot in
+  [#24940](https://github.com/google-gemini/gemini-cli/pull/24940)
+- feat(plan): require user confirmation for activate_skill in Plan Mode by
+  @ruomengz in [#24946](https://github.com/google-gemini/gemini-cli/pull/24946)
+- feat(test-utils): add CPU performance integration test harness by @sripasg in
+  [#24951](https://github.com/google-gemini/gemini-cli/pull/24951)
+- fix(cli-ui): enable Ctrl+Backspace for word deletion in Windows Terminal by
+  @dogukanozen in
+  [#21447](https://github.com/google-gemini/gemini-cli/pull/21447)
+- test(sdk): add unit tests for GeminiCliSession by @AdamyaSingh7 in
+  [#21897](https://github.com/google-gemini/gemini-cli/pull/21897)
+- fix(core): resolve windows symlink bypass and stabilize sandbox integration
+  tests by @ehedlund in
+  [#24834](https://github.com/google-gemini/gemini-cli/pull/24834)
+- fix(cli): restore file path display in edit and write tool confirmations by
+  @jwhelangoog in
+  [#24974](https://github.com/google-gemini/gemini-cli/pull/24974)
+- feat(core): refine shell tool description display logic by @jwhelangoog in
+  [#24903](https://github.com/google-gemini/gemini-cli/pull/24903)
+- fix(core): dynamic session ID injection to resolve resume bugs by @scidomino
+  in [#24972](https://github.com/google-gemini/gemini-cli/pull/24972)
+- Update ink version to 6.6.9 by @jacob314 in
+  [#24980](https://github.com/google-gemini/gemini-cli/pull/24980)
+- Generalize evals infra to support more types of evals, organization and
+  queuing of named suites by @gundermanc in
+  [#24941](https://github.com/google-gemini/gemini-cli/pull/24941)
+- fix(cli): optimize startup with lightweight parent process by @sehoon38 in
+  [#24667](https://github.com/google-gemini/gemini-cli/pull/24667)
+- refactor(sandbox): use centralized sandbox paths in macOS Seatbelt
+  implementation by @ehedlund in
+  [#24984](https://github.com/google-gemini/gemini-cli/pull/24984)
+- feat(cli): refine tool output formatting for compact mode by @jwhelangoog in
+  [#24677](https://github.com/google-gemini/gemini-cli/pull/24677)
+- fix(sdk): skip broken sendStream tests to unblock nightly by @SandyTao520 in
+  [#25000](https://github.com/google-gemini/gemini-cli/pull/25000)
+- refactor(core): use centralized path resolution for Linux sandbox by @ehedlund
+  in [#24985](https://github.com/google-gemini/gemini-cli/pull/24985)
+- Support ctrl+shift+g by @jacob314 in
+  [#25035](https://github.com/google-gemini/gemini-cli/pull/25035)
+- feat(core): refactor subagent tool to unified invoke_subagent tool by
+  @abhipatel12 in
+  [#24489](https://github.com/google-gemini/gemini-cli/pull/24489)
+- fix(core): add explicit git identity env vars to prevent sandbox checkpointing
+  error by @mrpmohiburrahman in
+  [#19775](https://github.com/google-gemini/gemini-cli/pull/19775)
+- fix: respect hideContextPercentage when FooterConfigDialog is closed without
+  changes by @chernistry in
+  [#24773](https://github.com/google-gemini/gemini-cli/pull/24773)
+- fix(cli): suppress unhandled AbortError logs during request cancellation by
+  @euxaristia in
+  [#22621](https://github.com/google-gemini/gemini-cli/pull/22621)
+- Automated documentation audit by @g-samroberts in
+  [#24567](https://github.com/google-gemini/gemini-cli/pull/24567)
+- feat(cli): implement useAgentStream hook by @mbleigh in
+  [#24292](https://github.com/google-gemini/gemini-cli/pull/24292)
+- refactor(plan) Clean default plan toml by @ruomengz in
+  [#25037](https://github.com/google-gemini/gemini-cli/pull/25037)
+- refactor(core): remove legacy subagent wrapping tools by @abhipatel12 in
+  [#25053](https://github.com/google-gemini/gemini-cli/pull/25053)
+- fix(core): honor retryDelay in RetryInfo for 503 errors by @yunaseoul in
+  [#25057](https://github.com/google-gemini/gemini-cli/pull/25057)
+- fix(core): remediate subagent memory leaks using AbortSignal in MessageBus by
+  @abhipatel12 in
+  [#25048](https://github.com/google-gemini/gemini-cli/pull/25048)
+- feat(cli): wire up useAgentStream in AppContainer by @mbleigh in
+  [#24297](https://github.com/google-gemini/gemini-cli/pull/24297)
+- feat(core): migrate chat recording to JSONL streaming by @spencer426 in
+  [#23749](https://github.com/google-gemini/gemini-cli/pull/23749)
+- fix(core): clear 5-minute timeouts in oauth flow to prevent memory leaks by
+  @spencer426 in
+  [#24968](https://github.com/google-gemini/gemini-cli/pull/24968)
+- fix(sandbox): centralize async git worktree resolution and enforce read-only
+  security by @ehedlund in
+  [#25040](https://github.com/google-gemini/gemini-cli/pull/25040)
+- feat(test): add high-volume shell test and refine perf harness by @sripasg in
+  [#24983](https://github.com/google-gemini/gemini-cli/pull/24983)
+- fix(core): silently handle EPERM when listing dir structure by @scidomino in
+  [#25066](https://github.com/google-gemini/gemini-cli/pull/25066)
+- Changelog for v0.37.1 by @gemini-cli-robot in
+  [#25055](https://github.com/google-gemini/gemini-cli/pull/25055)
+- fix: decode Uint8Array and multi-byte UTF-8 in API error messages by
+  @kimjune01 in [#23341](https://github.com/google-gemini/gemini-cli/pull/23341)
+- Automated documentation audit results by @g-samroberts in
+  [#22755](https://github.com/google-gemini/gemini-cli/pull/22755)
+- debugging(ui): add optional debugRainbow setting by @jacob314 in
+  [#25088](https://github.com/google-gemini/gemini-cli/pull/25088)
+- fix: resolve lifecycle memory leaks by cleaning up listeners and root closures
   by @spencer426 in
-  [#24566](https://github.com/google-gemini/gemini-cli/pull/24566)
-- Add an eval for and fix unsafe cloning behavior. by @gundermanc in
-  [#24457](https://github.com/google-gemini/gemini-cli/pull/24457)
-- fix(policy): allow complete_task in plan mode by @abhipatel12 in
-  [#24771](https://github.com/google-gemini/gemini-cli/pull/24771)
-- feat(telemetry): add browser agent clearcut metrics by @gsquared94 in
-  [#24688](https://github.com/google-gemini/gemini-cli/pull/24688)
-- feat(cli): support selective topic expansion and click-to-expand by
-  @Abhijit-2592 in
-  [#24793](https://github.com/google-gemini/gemini-cli/pull/24793)
-- temporarily disable sandbox integration test on windows by @ehedlund in
-  [#24786](https://github.com/google-gemini/gemini-cli/pull/24786)
-- Remove flakey test by @scidomino in
-  [#24837](https://github.com/google-gemini/gemini-cli/pull/24837)
-- Alisa/approve button by @alisa-alisa in
-  [#24645](https://github.com/google-gemini/gemini-cli/pull/24645)
-- feat(hooks): display hook system messages in UI by @mbleigh in
-  [#24616](https://github.com/google-gemini/gemini-cli/pull/24616)
-- fix(core): propagate BeforeModel hook model override end-to-end by @krishdef7
-  in [#24784](https://github.com/google-gemini/gemini-cli/pull/24784)
-- chore: fix formatting for behavioral eval skill reference file by @abhipatel12
-  in [#24846](https://github.com/google-gemini/gemini-cli/pull/24846)
-- fix: use directory junctions on Windows for skill linking by @enjoykumawat in
-  [#24823](https://github.com/google-gemini/gemini-cli/pull/24823)
-- fix(cli): prevent multiple banner increments on remount by @sehoon38 in
-  [#24843](https://github.com/google-gemini/gemini-cli/pull/24843)
-- feat(acp): add /help command by @sripasg in
-  [#24839](https://github.com/google-gemini/gemini-cli/pull/24839)
-- fix(core): remove tmux alternate buffer warning by @jackwotherspoon in
-  [#24852](https://github.com/google-gemini/gemini-cli/pull/24852)
-- Improve sandbox error matching and caching by @DavidAPierce in
-  [#24550](https://github.com/google-gemini/gemini-cli/pull/24550)
-- feat(core): add agent protocol UI types and experimental flag by @mbleigh in
-  [#24275](https://github.com/google-gemini/gemini-cli/pull/24275)
-- feat(core): use experiment flags for default fetch timeouts by @yunaseoul in
-  [#24261](https://github.com/google-gemini/gemini-cli/pull/24261)
-- Revert "fix(ui): improve narration suppression and reduce flicker (#2… by
+  [#25049](https://github.com/google-gemini/gemini-cli/pull/25049)
+- docs(cli): updates f12 description to be more precise by @JayadityaGit in
+  [#15816](https://github.com/google-gemini/gemini-cli/pull/15816)
+- fix(cli): mark /settings as unsafe to run concurrently by @jacob314 in
+  [#25061](https://github.com/google-gemini/gemini-cli/pull/25061)
+- fix(core): remove buffer slice to prevent OOM on large output streams by
+  @spencer426 in
+  [#25094](https://github.com/google-gemini/gemini-cli/pull/25094)
+- feat(core): persist subagent agentId in tool call records by @abhipatel12 in
+  [#25092](https://github.com/google-gemini/gemini-cli/pull/25092)
+- chore(core): increase codebase investigator turn limits to 50 by @abhipatel12
+  in [#25125](https://github.com/google-gemini/gemini-cli/pull/25125)
+- refactor(core): consolidate execute() arguments into ExecuteOptions by
+  @mbleigh in [#25101](https://github.com/google-gemini/gemini-cli/pull/25101)
+- feat(core): add Strategic Re-evaluation guidance to system prompt by
+  @aishaneeshah in
+  [#25062](https://github.com/google-gemini/gemini-cli/pull/25062)
+- fix(core): preserve shell execution config fields on update by
+  @jasonmatthewsuhari in
+  [#25113](https://github.com/google-gemini/gemini-cli/pull/25113)
+- docs: add vi shortcuts and clarify MCP sandbox setup by @chrisjcthomas in
+  [#21679](https://github.com/google-gemini/gemini-cli/pull/21679)
+- fix(cli): pass session id to interactive shell executions by
+  @jasonmatthewsuhari in
+  [#25114](https://github.com/google-gemini/gemini-cli/pull/25114)
+- fix(cli): resolve text sanitization data loss due to C1 control characters by
+  @euxaristia in
+  [#22624](https://github.com/google-gemini/gemini-cli/pull/22624)
+- feat(core): add large memory regression test by @cynthialong0-0 in
+  [#25059](https://github.com/google-gemini/gemini-cli/pull/25059)
+- fix(core): resolve PTY exhaustion and orphan MCP subprocess leaks by
+  @spencer426 in
+  [#25079](https://github.com/google-gemini/gemini-cli/pull/25079)
+- chore(deps): update vulnerable dependencies via npm audit fix by @scidomino in
+  [#25140](https://github.com/google-gemini/gemini-cli/pull/25140)
+- perf(sandbox): optimize Windows sandbox initialization via native ACL
+  application by @ehedlund in
+  [#25077](https://github.com/google-gemini/gemini-cli/pull/25077)
+- chore: switch from keytar to @github/keytar by @cocosheng-g in
+  [#25143](https://github.com/google-gemini/gemini-cli/pull/25143)
+- fix: improve audio MIME normalization and validation in file reads by
+  @junaiddshaukat in
+  [#21636](https://github.com/google-gemini/gemini-cli/pull/21636)
+- docs: Update docs-audit to include changes in PR body by @g-samroberts in
+  [#25153](https://github.com/google-gemini/gemini-cli/pull/25153)
+- docs: correct documentation for enforced authentication type by @cocosheng-g
+  in [#25142](https://github.com/google-gemini/gemini-cli/pull/25142)
+- fix(cli): exclude update_topic from confirmation queue count by @Abhijit-2592
+  in [#24945](https://github.com/google-gemini/gemini-cli/pull/24945)
+- Memory fix for trace's streamWrapper. by @anthraxmilkshake in
+  [#25089](https://github.com/google-gemini/gemini-cli/pull/25089)
+- fix(core): fix quota footer for non-auto models and improve display by
+  @jackwotherspoon in
+  [#25121](https://github.com/google-gemini/gemini-cli/pull/25121)
+- docs(contributing): clarify self-assignment policy for issues by @jmr in
+  [#23087](https://github.com/google-gemini/gemini-cli/pull/23087)
+- feat(core): add skill patching support with /memory inbox integration by
+  @SandyTao520 in
+  [#25148](https://github.com/google-gemini/gemini-cli/pull/25148)
+- Stop suppressing thoughts and text in model response by @gundermanc in
+  [#25073](https://github.com/google-gemini/gemini-cli/pull/25073)
+- fix(release): prefix git hash in nightly versions to prevent semver
+  normalization by @SandyTao520 in
+  [#25304](https://github.com/google-gemini/gemini-cli/pull/25304)
+- feat(cli): extract QuotaContext and resolve infinite render loop by @Adib234
+  in [#24959](https://github.com/google-gemini/gemini-cli/pull/24959)
+- refactor(core): extract and centralize sandbox path utilities by @ehedlund in
+  [#25305](https://github.com/google-gemini/gemini-cli/pull/25305)
+- feat(ui): added enhancements to scroll momentum by @devr0306 in
+  [#24447](https://github.com/google-gemini/gemini-cli/pull/24447)
+- fix(core): replace custom binary detection with isbinaryfile to correctly
+  handle UTF-8 (U+FFFD) by @Anjaligarhwal in
+  [#25297](https://github.com/google-gemini/gemini-cli/pull/25297)
+- feat(agent): implement tool-controlled display protocol (Steps 2-3) by
+  @mbleigh in [#25134](https://github.com/google-gemini/gemini-cli/pull/25134)
+- Stop showing scrollbar unless we are in terminalBuffer mode by @jacob314 in
+  [#25320](https://github.com/google-gemini/gemini-cli/pull/25320)
+- feat: support auth block in MCP servers config in agents by @TanmayVartak in
+  [#24770](https://github.com/google-gemini/gemini-cli/pull/24770)
+- fix(core): expose GEMINI_PLANS_DIR to hook environment by @Adib234 in
+  [#25296](https://github.com/google-gemini/gemini-cli/pull/25296)
+- feat(core): implement silent fallback for Plan Mode model routing by @jerop in
+  [#25317](https://github.com/google-gemini/gemini-cli/pull/25317)
+- fix: correct redirect count increment in fetchJson by @KevinZhao in
+  [#24896](https://github.com/google-gemini/gemini-cli/pull/24896)
+- fix(core): prevent secondary crash in ModelRouterService finally block by
   @gundermanc in
-  [#24857](https://github.com/google-gemini/gemini-cli/pull/24857)
-- refactor(cli): remove duplication in interactive shell awaiting input hint by
-  @JayadityaGit in
-  [#24801](https://github.com/google-gemini/gemini-cli/pull/24801)
-- refactor(core): make LegacyAgentSession dependencies optional by @mbleigh in
-  [#24287](https://github.com/google-gemini/gemini-cli/pull/24287)
-- Changelog for v0.37.0-preview.2 by @gemini-cli-robot in
-  [#24848](https://github.com/google-gemini/gemini-cli/pull/24848)
-- fix(cli): always show shell command description or actual command by @jacob314
-  in [#24774](https://github.com/google-gemini/gemini-cli/pull/24774)
-- Added flag for ept size and increased default size by @devr0306 in
-  [#24859](https://github.com/google-gemini/gemini-cli/pull/24859)
-- fix(core): dispose Scheduler to prevent McpProgress listener leak by
-  @Anjaligarhwal in
-  [#24870](https://github.com/google-gemini/gemini-cli/pull/24870)
-- fix(cli): switch default back to terminalBuffer=false and fix regressions
-  introduced for that mode by @jacob314 in
-  [#24873](https://github.com/google-gemini/gemini-cli/pull/24873)
-- feat(cli): switch to ctrl+g from ctrl-x by @jacob314 in
-  [#24861](https://github.com/google-gemini/gemini-cli/pull/24861)
-- fix: isolate concurrent browser agent instances by @gsquared94 in
-  [#24794](https://github.com/google-gemini/gemini-cli/pull/24794)
-- docs: update MCP server OAuth redirect port documentation by @adamfweidman in
-  [#24844](https://github.com/google-gemini/gemini-cli/pull/24844)
+  [#25333](https://github.com/google-gemini/gemini-cli/pull/25333)
+- feat(core): introduce decoupled ContextManager and Sidecar architecture by
+  @joshualitt in
+  [#24752](https://github.com/google-gemini/gemini-cli/pull/24752)
+- docs(core): update generalist agent documentation by @abhipatel12 in
+  [#25325](https://github.com/google-gemini/gemini-cli/pull/25325)
+- chore(mcp): check MCP error code over brittle string match by @jackwotherspoon
+  in [#25381](https://github.com/google-gemini/gemini-cli/pull/25381)
+- feat(plan): update plan mode prompt to allow showing plan content by @ruomengz
+  in [#25058](https://github.com/google-gemini/gemini-cli/pull/25058)
+- test(core): improve sandbox integration test coverage and fix OS-specific
+  failures by @ehedlund in
+  [#25307](https://github.com/google-gemini/gemini-cli/pull/25307)
+- fix(core): use debug level for keychain fallback logging by @ehedlund in
+  [#25398](https://github.com/google-gemini/gemini-cli/pull/25398)
+- feat(test): add a performance test in asian language by @cynthialong0-0 in
+  [#25392](https://github.com/google-gemini/gemini-cli/pull/25392)
+- feat(cli): enable mouse clicking for cursor positioning in AskUser multi-line
+  answers by @Adib234 in
+  [#24630](https://github.com/google-gemini/gemini-cli/pull/24630)
+- fix(core): detect kmscon terminal as supporting true color by @claygeo in
+  [#25282](https://github.com/google-gemini/gemini-cli/pull/25282)
+- ci: add agent session drift check workflow by @adamfweidman in
+  [#25389](https://github.com/google-gemini/gemini-cli/pull/25389)
+- use macos-latest-large runner where applicable. by @scidomino in
+  [#25413](https://github.com/google-gemini/gemini-cli/pull/25413)
+- Changelog for v0.37.2 by @gemini-cli-robot in
+  [#25336](https://github.com/google-gemini/gemini-cli/pull/25336)
 
 **Full Changelog**:
-https://github.com/google-gemini/gemini-cli/compare/v0.37.0-preview.2...v0.38.0-preview.0
+https://github.com/google-gemini/gemini-cli/compare/v0.38.0-preview.0...v0.39.0-preview.0

docs/reference/configuration.md

@@ -2148,6 +2148,21 @@ the `advanced.excludedEnvVars` setting in your `settings.json` file.
   - When set, overrides the default API version used by the SDK.
   - Example: `export GOOGLE_GENAI_API_VERSION="v1"` (Windows PowerShell:
     `$env:GOOGLE_GENAI_API_VERSION="v1"`)
+- **`GOOGLE_GEMINI_BASE_URL`**:
+  - Overrides the default base URL for Gemini API requests (when using
+    `gemini-api-key` authentication).
+  - Must be a valid URL. For security, it must use HTTPS unless pointing to
+    `localhost` (or `127.0.0.1` / `[::1]`).
+  - Example: `export GOOGLE_GEMINI_BASE_URL="https://my-proxy.com"` (Windows
+    PowerShell: `$env:GOOGLE_GEMINI_BASE_URL="https://my-proxy.com"`)
+- **`GOOGLE_VERTEX_BASE_URL`**:
+  - Overrides the default base URL for Vertex AI API requests (when using
+    `vertex-ai` authentication).
+  - Must be a valid URL. For security, it must use HTTPS unless pointing to
+    `localhost` (or `127.0.0.1` / `[::1]`).
+  - Example: `export GOOGLE_VERTEX_BASE_URL="https://my-vertex-proxy.com"`
+    (Windows PowerShell:
+    `$env:GOOGLE_VERTEX_BASE_URL="https://my-vertex-proxy.com"`)
 - **`OTLP_GOOGLE_CLOUD_PROJECT`**:
   - Your Google Cloud Project ID for Telemetry in Google Cloud
   - Example: `export OTLP_GOOGLE_CLOUD_PROJECT="YOUR_PROJECT_ID"` (Windows

evals/cli_help_delegation.eval.ts

@@ -17,9 +17,17 @@ describe('CliHelpAgent Delegation', () => {
     timeout: 60000,
     assert: async (rig, _result) => {
       const toolLogs = rig.readToolLogs();
-      const toolCallIndex = toolLogs.findIndex(
-        (log) => log.toolRequest.name === 'cli_help',
-      );
+      const toolCallIndex = toolLogs.findIndex((log) => {
+        if (log.toolRequest.name === 'invoke_agent') {
+          try {
+            const args = JSON.parse(log.toolRequest.args);
+            return args.agent_name === 'cli_help';
+          } catch {
+            return false;
+          }
+        }
+        return false;
+      });
       expect(toolCallIndex).toBeGreaterThan(-1);
       expect(toolCallIndex).toBeLessThan(5); // Called within first 5 turns
     },

evals/component-test-helper.ts

@@ -16,6 +16,7 @@ import fs from 'node:fs';
 import path from 'node:path';
 import os from 'node:os';
 import { randomUUID } from 'node:crypto';
+import { vi } from 'vitest';
 import {
   Config,
   type ConfigParameters,
@@ -52,6 +53,7 @@ export interface ComponentEvalCase extends BaseEvalCase {
 export class ComponentRig {
   public config: Config | undefined;
   public testDir: string;
+  public homeDir: string;
   public sessionId: string;
 
   constructor(
@@ -61,6 +63,9 @@ export class ComponentRig {
     this.testDir = fs.mkdtempSync(
       path.join(os.tmpdir(), `gemini-component-rig-${uniqueId.slice(0, 8)}-`),
     );
+    this.homeDir = fs.mkdtempSync(
+      path.join(os.tmpdir(), `gemini-component-home-${uniqueId.slice(0, 8)}-`),
+    );
     this.sessionId = `test-session-${uniqueId}`;
   }
 
@@ -89,12 +94,23 @@ export class ComponentRig {
     this.config = makeFakeConfig(configParams);
     await this.config.initialize();
 
-    // Refresh auth using USE_GEMINI to initialize the real BaseLlmClient
+    // Refresh auth using USE_GEMINI to initialize the real BaseLlmClient.
+    // This must happen BEFORE stubbing GEMINI_CLI_HOME because OAuth credential
+    // lookup resolves through homedir() → GEMINI_CLI_HOME.
     await this.config.refreshAuth(AuthType.USE_GEMINI);
+
+    // Isolate storage paths (session files, skills, extraction state) by
+    // pointing GEMINI_CLI_HOME at a per-test temp directory.  Storage resolves
+    // global paths through `homedir()` which reads this env var.  This is set
+    // after auth so credential lookup uses the real home directory.
+    vi.stubEnv('GEMINI_CLI_HOME', this.homeDir);
   }
 
   async cleanup() {
+    await this.config?.dispose();
+    vi.unstubAllEnvs();
     fs.rmSync(this.testDir, { recursive: true, force: true });
+    fs.rmSync(this.homeDir, { recursive: true, force: true });
   }
 }
 

evals/generalist_agent.eval.ts

@@ -26,11 +26,22 @@ describe('generalist_agent', () => {
     prompt:
       'Please use the generalist agent to create a file called "generalist_test_file.txt" containing exactly the following text: success',
     assert: async (rig) => {
-      // 1) Verify the generalist agent was invoked
-      const foundToolCall = await rig.waitForToolCall('generalist');
+      // 1) Verify the generalist agent was invoked via invoke_agent
+      const foundToolCall = await rig.waitForToolCall(
+        'invoke_agent',
+        undefined,
+        (args) => {
+          try {
+            const parsed = JSON.parse(args);
+            return parsed.agent_name === 'generalist';
+          } catch {
+            return false;
+          }
+        },
+      );
       expect(
         foundToolCall,
-        'Expected to find a tool call for generalist agent',
+        'Expected to find an invoke_agent tool call for generalist agent',
       ).toBeTruthy();
 
       // 2) Verify the file was created as expected

evals/save_memory.eval.ts

@@ -145,22 +145,30 @@ describe('save_memory', () => {
     },
   });
 
-  const ignoringDbSchemaLocation =
-    "Agent ignores workspace's database schema location";
+  const savingDbSchemaLocationAsProjectMemory =
+    'Agent saves workspace database schema location as project memory';
   evalTest('USUALLY_PASSES', {
     suiteName: 'default',
     suiteType: 'behavioral',
-    name: ignoringDbSchemaLocation,
+    name: savingDbSchemaLocationAsProjectMemory,
     prompt: `The database schema for this workspace is located in \`db/schema.sql\`.`,
     assert: async (rig, result) => {
-      await rig.waitForTelemetryReady();
-      const wasToolCalled = rig
-        .readToolLogs()
-        .some((log) => log.toolRequest.name === 'save_memory');
+      const wasToolCalled = await rig.waitForToolCall(
+        'save_memory',
+        undefined,
+        (args) => {
+          try {
+            const params = JSON.parse(args);
+            return params.scope === 'project';
+          } catch {
+            return false;
+          }
+        },
+      );
       expect(
         wasToolCalled,
-        'save_memory should not be called for workspace-specific information',
-      ).toBe(false);
+        'Expected save_memory to be called with scope="project" for workspace-specific information',
+      ).toBe(true);
 
       assertModelHasOutput(result);
     },
@@ -188,42 +196,59 @@ describe('save_memory', () => {
     },
   });
 
-  const ignoringBuildArtifactLocation =
-    'Agent ignores workspace build artifact location';
+  const savingBuildArtifactLocationAsProjectMemory =
+    'Agent saves workspace build artifact location as project memory';
   evalTest('USUALLY_PASSES', {
     suiteName: 'default',
     suiteType: 'behavioral',
-    name: ignoringBuildArtifactLocation,
+    name: savingBuildArtifactLocationAsProjectMemory,
     prompt: `In this workspace, build artifacts are stored in the \`dist/artifacts\` directory.`,
     assert: async (rig, result) => {
-      await rig.waitForTelemetryReady();
-      const wasToolCalled = rig
-        .readToolLogs()
-        .some((log) => log.toolRequest.name === 'save_memory');
+      const wasToolCalled = await rig.waitForToolCall(
+        'save_memory',
+        undefined,
+        (args) => {
+          try {
+            const params = JSON.parse(args);
+            return params.scope === 'project';
+          } catch {
+            return false;
+          }
+        },
+      );
       expect(
         wasToolCalled,
-        'save_memory should not be called for workspace-specific information',
-      ).toBe(false);
+        'Expected save_memory to be called with scope="project" for workspace-specific information',
+      ).toBe(true);
 
       assertModelHasOutput(result);
     },
   });
 
-  const ignoringMainEntryPoint = "Agent ignores workspace's main entry point";
+  const savingMainEntryPointAsProjectMemory =
+    'Agent saves workspace main entry point as project memory';
   evalTest('USUALLY_PASSES', {
     suiteName: 'default',
     suiteType: 'behavioral',
-    name: ignoringMainEntryPoint,
+    name: savingMainEntryPointAsProjectMemory,
     prompt: `The main entry point for this workspace is \`src/index.js\`.`,
     assert: async (rig, result) => {
-      await rig.waitForTelemetryReady();
-      const wasToolCalled = rig
-        .readToolLogs()
-        .some((log) => log.toolRequest.name === 'save_memory');
+      const wasToolCalled = await rig.waitForToolCall(
+        'save_memory',
+        undefined,
+        (args) => {
+          try {
+            const params = JSON.parse(args);
+            return params.scope === 'project';
+          } catch {
+            return false;
+          }
+        },
+      );
       expect(
         wasToolCalled,
-        'save_memory should not be called for workspace-specific information',
-      ).toBe(false);
+        'Expected save_memory to be called with scope="project" for workspace-specific information',
+      ).toBe(true);
 
       assertModelHasOutput(result);
     },
@@ -317,13 +342,13 @@ describe('save_memory', () => {
       'Please save any persistent preferences or facts about me from our conversation to memory.',
     assert: async (rig, result) => {
       const wasToolCalled = await rig.waitForToolCall(
-        'save_memory',
+        'invoke_agent',
         undefined,
-        (args) => /vitest/i.test(args),
+        (args) => /save_memory/i.test(args) && /vitest/i.test(args),
       );
       expect(
         wasToolCalled,
-        'Expected save_memory to be called with the Vitest preference from the conversation history',
+        'Expected invoke_agent to be called with save_memory agent and the Vitest preference from the conversation history',
       ).toBe(true);
 
       assertModelHasOutput(result);
@@ -379,8 +404,15 @@ describe('save_memory', () => {
     ],
     prompt: 'Please save the preferences I mentioned earlier to memory.',
     assert: async (rig, result) => {
-      const wasToolCalled = await rig.waitForToolCall('save_memory');
-      expect(wasToolCalled, 'Expected save_memory to be called').toBe(true);
+      const wasToolCalled = await rig.waitForToolCall(
+        'invoke_agent',
+        undefined,
+        (args) => /save_memory/i.test(args),
+      );
+      expect(
+        wasToolCalled,
+        'Expected invoke_agent to be called with save_memory agent',
+      ).toBe(true);
 
       assertModelHasOutput(result);
     },

evals/skill_extraction.eval.ts

@@ -0,0 +1,349 @@
+/**
+ * @license
+ * Copyright 2026 Google LLC
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+import fsp from 'node:fs/promises';
+import path from 'node:path';
+import { describe, expect } from 'vitest';
+import {
+  type Config,
+  ApprovalMode,
+  SESSION_FILE_PREFIX,
+  getProjectHash,
+  startMemoryService,
+} from '@google/gemini-cli-core';
+import { componentEvalTest } from './component-test-helper.js';
+
+interface SeedSession {
+  sessionId: string;
+  summary: string;
+  userTurns: string[];
+  timestampOffsetMinutes: number;
+}
+
+interface MessageRecord {
+  id: string;
+  timestamp: string;
+  type: string;
+  content: Array<{ text: string }>;
+}
+
+const WORKSPACE_FILES = {
+  'package.json': JSON.stringify(
+    {
+      name: 'skill-extraction-eval',
+      private: true,
+      scripts: {
+        build: 'echo build',
+        lint: 'echo lint',
+        test: 'echo test',
+      },
+    },
+    null,
+    2,
+  ),
+  'README.md': `# Skill Extraction Eval
+
+This workspace exists to exercise background skill extraction from prior chats.
+`,
+};
+
+function buildMessages(userTurns: string[]): MessageRecord[] {
+  const baseTime = new Date(Date.now() - 6 * 60 * 60 * 1000).toISOString();
+  return userTurns.flatMap((text, index) => [
+    {
+      id: `u${index + 1}`,
+      timestamp: baseTime,
+      type: 'user',
+      content: [{ text }],
+    },
+    {
+      id: `a${index + 1}`,
+      timestamp: baseTime,
+      type: 'gemini',
+      content: [{ text: `Acknowledged: ${index + 1}` }],
+    },
+  ]);
+}
+
+async function seedSessions(
+  config: Config,
+  sessions: SeedSession[],
+): Promise<void> {
+  const chatsDir = path.join(config.storage.getProjectTempDir(), 'chats');
+  await fsp.mkdir(chatsDir, { recursive: true });
+
+  const projectRoot = config.storage.getProjectRoot();
+
+  for (const session of sessions) {
+    const timestamp = new Date(
+      Date.now() - session.timestampOffsetMinutes * 60 * 1000,
+    )
+      .toISOString()
+      .slice(0, 16)
+      .replace(/:/g, '-');
+    const filename = `${SESSION_FILE_PREFIX}${timestamp}-${session.sessionId.slice(0, 8)}.json`;
+    const conversation = {
+      sessionId: session.sessionId,
+      projectHash: getProjectHash(projectRoot),
+      summary: session.summary,
+      startTime: new Date(Date.now() - 7 * 60 * 60 * 1000).toISOString(),
+      lastUpdated: new Date(Date.now() - 4 * 60 * 60 * 1000).toISOString(),
+      messages: buildMessages(session.userTurns),
+    };
+
+    await fsp.writeFile(
+      path.join(chatsDir, filename),
+      JSON.stringify(conversation, null, 2),
+    );
+  }
+}
+
+async function runExtractionAndReadState(config: Config): Promise<{
+  state: { runs: Array<{ sessionIds: string[]; skillsCreated: string[] }> };
+  skillsDir: string;
+}> {
+  await startMemoryService(config);
+
+  const memoryDir = config.storage.getProjectMemoryTempDir();
+  const skillsDir = config.storage.getProjectSkillsMemoryDir();
+  const statePath = path.join(memoryDir, '.extraction-state.json');
+
+  const raw = await fsp.readFile(statePath, 'utf-8');
+  const state = JSON.parse(raw) as {
+    runs?: Array<{ sessionIds?: string[]; skillsCreated?: string[] }>;
+  };
+  if (!Array.isArray(state.runs) || state.runs.length === 0) {
+    throw new Error('Skill extraction finished without writing any run state');
+  }
+
+  return {
+    state: {
+      runs: state.runs.map((run) => ({
+        sessionIds: Array.isArray(run.sessionIds) ? run.sessionIds : [],
+        skillsCreated: Array.isArray(run.skillsCreated)
+          ? run.skillsCreated
+          : [],
+      })),
+    },
+    skillsDir,
+  };
+}
+
+async function readSkillBodies(skillsDir: string): Promise<string[]> {
+  try {
+    const entries = await fsp.readdir(skillsDir, { withFileTypes: true });
+    const skillDirs = entries.filter((entry) => entry.isDirectory());
+    const bodies = await Promise.all(
+      skillDirs.map((entry) =>
+        fsp.readFile(path.join(skillsDir, entry.name, 'SKILL.md'), 'utf-8'),
+      ),
+    );
+    return bodies;
+  } catch {
+    return [];
+  }
+}
+
+/**
+ * Shared configOverrides for all skill extraction component evals.
+ * - experimentalMemoryManager: enables the memory extraction pipeline.
+ * - approvalMode: YOLO auto-approves tool calls (write_file, read_file) so the
+ *   background agent can execute without interactive confirmation.
+ */
+const EXTRACTION_CONFIG_OVERRIDES = {
+  experimentalMemoryManager: true,
+  approvalMode: ApprovalMode.YOLO,
+};
+
+describe('Skill Extraction', () => {
+  componentEvalTest('USUALLY_PASSES', {
+    suiteName: 'skill-extraction',
+    suiteType: 'component-level',
+    name: 'ignores one-off incidents even when session summaries look similar',
+    files: WORKSPACE_FILES,
+    timeout: 180000,
+    configOverrides: EXTRACTION_CONFIG_OVERRIDES,
+    setup: async (config) => {
+      await seedSessions(config, [
+        {
+          sessionId: 'incident-login-redirect',
+          summary: 'Debug login redirect loop in staging',
+          timestampOffsetMinutes: 420,
+          userTurns: [
+            'We only need a one-off fix for incident INC-4412 on branch hotfix/login-loop.',
+            'The exact failing string is ERR_REDIRECT_4412 and this workaround is incident-specific.',
+            'Patch packages/auth/src/redirect.ts just for this branch and do not generalize it.',
+            'The thing that worked was deleting the stale staging cookie before retrying.',
+            'This is not a normal workflow and should not become a reusable instruction.',
+            'It only reproduced against the 2026-04-08 staging rollout.',
+            'After the cookie clear, the branch-specific redirect logic passed.',
+            'Do not turn this incident writeup into a standing process.',
+            'Yes, the hotfix worked for this exact redirect-loop incident.',
+            'Close out INC-4412 once the staging login succeeds again.',
+          ],
+        },
+        {
+          sessionId: 'incident-login-timeout',
+          summary: 'Debug login callback timeout in staging',
+          timestampOffsetMinutes: 360,
+          userTurns: [
+            'This is another one-off staging incident, this time TICKET-991 for callback timeout.',
+            'The exact failing string is ERR_CALLBACK_TIMEOUT_991 and it is unrelated to the redirect loop.',
+            'The temporary fix was rotating the staging secret and deleting a bad feature-flag row.',
+            'Do not write a generic login-debugging playbook from this.',
+            'This only applied to the callback timeout during the April rollout.',
+            'The successful fix was specific to the stale secret in staging.',
+            'It does not define a durable repo workflow for future tasks.',
+            'After rotating the secret, the callback timeout stopped reproducing.',
+            'Treat this as incident response only, not a reusable skill.',
+            'Once staging passed again, we closed TICKET-991.',
+          ],
+        },
+      ]);
+    },
+    assert: async (config) => {
+      const { state, skillsDir } = await runExtractionAndReadState(config);
+      const skillBodies = await readSkillBodies(skillsDir);
+
+      expect(state.runs).toHaveLength(1);
+      expect(state.runs[0].sessionIds).toHaveLength(2);
+      expect(state.runs[0].skillsCreated).toEqual([]);
+      expect(skillBodies).toEqual([]);
+    },
+  });
+
+  componentEvalTest('USUALLY_PASSES', {
+    suiteName: 'skill-extraction',
+    suiteType: 'component-level',
+    name: 'extracts a repeated project-specific workflow into a skill',
+    files: WORKSPACE_FILES,
+    timeout: 180000,
+    configOverrides: EXTRACTION_CONFIG_OVERRIDES,
+    setup: async (config) => {
+      await seedSessions(config, [
+        {
+          sessionId: 'settings-docs-regen-1',
+          summary: 'Update settings docs after adding a config option',
+          timestampOffsetMinutes: 420,
+          userTurns: [
+            'When we add a new config option, we have to regenerate the settings docs in a specific order.',
+            'The sequence that worked was npm run predocs:settings, npm run schema:settings, then npm run docs:settings.',
+            'Do not hand-edit generated settings docs.',
+            'If predocs is skipped, the generated schema docs miss the new defaults.',
+            'Update the source first, then run that generation sequence.',
+            'After regenerating, verify the schema output and docs changed together.',
+            'We used this same sequence the last time we touched settings docs.',
+            'That ordered workflow passed and produced the expected generated files.',
+            'Please keep the exact command order because reversing it breaks the output.',
+            'Yes, the generated settings docs were correct after those three commands.',
+          ],
+        },
+        {
+          sessionId: 'settings-docs-regen-2',
+          summary: 'Regenerate settings schema docs for another new setting',
+          timestampOffsetMinutes: 360,
+          userTurns: [
+            'We are touching another setting, so follow the same settings-doc regeneration workflow again.',
+            'Run npm run predocs:settings before npm run schema:settings and npm run docs:settings.',
+            'The project keeps generated settings docs in sync through those commands, not manual edits.',
+            'Skipping predocs caused stale defaults in the generated output before.',
+            'Change the source, then execute the same three commands in order.',
+            'Verify both the schema artifact and docs update together after regeneration.',
+            'This is the recurring workflow we use whenever a setting changes.',
+            'The exact order worked again on this second settings update.',
+            'Please preserve that ordering constraint for future settings changes.',
+            'Confirmed: the settings docs regenerated correctly with the same command sequence.',
+          ],
+        },
+      ]);
+    },
+    assert: async (config) => {
+      const { state, skillsDir } = await runExtractionAndReadState(config);
+      const skillBodies = await readSkillBodies(skillsDir);
+      const combinedSkills = skillBodies.join('\n\n');
+
+      expect(state.runs).toHaveLength(1);
+      expect(state.runs[0].sessionIds).toHaveLength(2);
+      expect(state.runs[0].skillsCreated.length).toBeGreaterThanOrEqual(1);
+      expect(skillBodies.length).toBeGreaterThanOrEqual(1);
+      expect(combinedSkills).toContain('npm run predocs:settings');
+      expect(combinedSkills).toContain('npm run schema:settings');
+      expect(combinedSkills).toContain('npm run docs:settings');
+      expect(combinedSkills).toMatch(/Verification/i);
+
+      // Verify the extraction agent activated skill-creator for design guidance.
+      expect(config.getSkillManager().isSkillActive('skill-creator')).toBe(
+        true,
+      );
+    },
+  });
+
+  componentEvalTest('USUALLY_PASSES', {
+    suiteName: 'skill-extraction',
+    suiteType: 'component-level',
+    name: 'extracts a repeated multi-step migration workflow with ordering constraints',
+    files: WORKSPACE_FILES,
+    timeout: 180000,
+    configOverrides: EXTRACTION_CONFIG_OVERRIDES,
+    setup: async (config) => {
+      await seedSessions(config, [
+        {
+          sessionId: 'db-migration-v12',
+          summary: 'Run database migration for v12 schema update',
+          timestampOffsetMinutes: 420,
+          userTurns: [
+            'Every time we change the database schema we follow a specific migration workflow.',
+            'First run npm run db:check to verify no pending migrations conflict.',
+            'Then run npm run db:migrate to apply the new migration files.',
+            'After migration, always run npm run db:validate to confirm schema integrity.',
+            'If db:validate fails, immediately run npm run db:rollback before anything else.',
+            'Never skip db:check — last time we did, two migrations collided and corrupted the index.',
+            'The ordering is critical: check, migrate, validate. Reversing migrate and validate caused silent data loss before.',
+            'This v12 migration passed after following that exact sequence.',
+            'We use this same three-step workflow every time the schema changes.',
+            'Confirmed: db:check, db:migrate, db:validate completed successfully for v12.',
+          ],
+        },
+        {
+          sessionId: 'db-migration-v13',
+          summary: 'Run database migration for v13 schema update',
+          timestampOffsetMinutes: 360,
+          userTurns: [
+            'New schema change for v13, following the same database migration workflow as before.',
+            'Start with npm run db:check to ensure no conflicting pending migrations.',
+            'Then npm run db:migrate to apply the v13 migration files.',
+            'Then npm run db:validate to confirm the schema is consistent.',
+            'If validation fails, run npm run db:rollback immediately — do not attempt manual fixes.',
+            'We learned the hard way that skipping db:check causes index corruption.',
+            'The check-migrate-validate order is mandatory for every schema change.',
+            'This is the same recurring workflow we used for v12 and earlier migrations.',
+            'The v13 migration passed with the same three-step sequence.',
+            'Confirmed: the standard db migration workflow succeeded again for v13.',
+          ],
+        },
+      ]);
+    },
+    assert: async (config) => {
+      const { state, skillsDir } = await runExtractionAndReadState(config);
+      const skillBodies = await readSkillBodies(skillsDir);
+      const combinedSkills = skillBodies.join('\n\n');
+
+      expect(state.runs).toHaveLength(1);
+      expect(state.runs[0].sessionIds).toHaveLength(2);
+      expect(state.runs[0].skillsCreated.length).toBeGreaterThanOrEqual(1);
+      expect(skillBodies.length).toBeGreaterThanOrEqual(1);
+      expect(combinedSkills).toContain('npm run db:check');
+      expect(combinedSkills).toContain('npm run db:migrate');
+      expect(combinedSkills).toContain('npm run db:validate');
+      expect(combinedSkills).toMatch(/rollback/i);
+
+      // Verify the extraction agent activated skill-creator for design guidance.
+      expect(config.getSkillManager().isSkillActive('skill-creator')).toBe(
+        true,
+      );
+    },
+  });
+});

evals/subtask_delegation.eval.ts

@@ -0,0 +1,130 @@
+/**
+ * @license
+ * Copyright 2026 Google LLC
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+import { describe, expect } from 'vitest';
+import {
+  TRACKER_CREATE_TASK_TOOL_NAME,
+  TRACKER_UPDATE_TASK_TOOL_NAME,
+} from '@google/gemini-cli-core';
+import { evalTest, TEST_AGENTS } from './test-helper.js';
+
+describe('subtask delegation eval test cases', () => {
+  /**
+   * Checks that the main agent can correctly decompose a complex, sequential
+   * task into subtasks using the task tracker and delegate each to the appropriate expert subagent.
+   *
+   * The task requires:
+   * 1. Reading requirements (researcher)
+   * 2. Implementing logic (developer)
+   * 3. Documenting (doc expert)
+   */
+  evalTest('USUALLY_PASSES', {
+    name: 'should delegate sequential subtasks to relevant experts using the task tracker',
+    params: {
+      settings: {
+        experimental: {
+          enableAgents: true,
+          taskTracker: true,
+        },
+      },
+    },
+    prompt:
+      'Please read the requirements in requirements.txt using a researcher, then implement the requested logic in src/logic.ts using a developer, and finally document the implementation in docs/logic.md using a documentation expert.',
+    files: {
+      '.gemini/agents/researcher.md': `---
+name: researcher
+description: Expert in reading files and extracting requirements.
+tools:
+  - read_file
+---
+You are the researcher. Read the provided file and extract requirements.`,
+      '.gemini/agents/developer.md': `---
+name: developer
+description: Expert in implementing logic in TypeScript.
+tools:
+  - write_file
+---
+You are the developer. Implement the requested logic in the specified file.`,
+      '.gemini/agents/doc-expert.md': `---
+name: doc-expert
+description: Expert in writing technical documentation.
+tools:
+  - write_file
+---
+You are the doc expert. Document the provided implementation clearly.`,
+      'requirements.txt':
+        'Implement a function named "calculateSum" that adds two numbers.',
+    },
+    assert: async (rig, _result) => {
+      // Verify tracker tasks were created
+      const wasCreateCalled = await rig.waitForToolCall(
+        TRACKER_CREATE_TASK_TOOL_NAME,
+      );
+      expect(wasCreateCalled).toBe(true);
+
+      const toolLogs = rig.readToolLogs();
+      const createCalls = toolLogs.filter(
+        (l) => l.toolRequest.name === TRACKER_CREATE_TASK_TOOL_NAME,
+      );
+      expect(createCalls.length).toBeGreaterThanOrEqual(3);
+
+      await rig.expectToolCallSuccess([
+        'researcher',
+        'developer',
+        'doc-expert',
+      ]);
+
+      const logicFile = rig.readFile('src/logic.ts');
+      const docFile = rig.readFile('docs/logic.md');
+
+      expect(logicFile).toContain('calculateSum');
+      expect(docFile).toBeTruthy();
+    },
+  });
+
+  /**
+   * Checks that the main agent can delegate a batch of independent subtasks
+   * to multiple subagents in parallel using the task tracker to manage state.
+   */
+  evalTest('USUALLY_PASSES', {
+    name: 'should delegate independent subtasks to specialists using the task tracker',
+    params: {
+      settings: {
+        experimental: {
+          enableAgents: true,
+          taskTracker: true,
+        },
+      },
+    },
+    prompt:
+      'Please update the project for internationalization (i18n), audit the security of the current code, and update the CSS to use a blue theme. Use specialized experts for each task.',
+    files: {
+      ...TEST_AGENTS.I18N_AGENT.asFile(),
+      ...TEST_AGENTS.SECURITY_AGENT.asFile(),
+      ...TEST_AGENTS.CSS_AGENT.asFile(),
+      'index.ts': 'console.log("Hello World");',
+    },
+    assert: async (rig, _result) => {
+      // Verify tracker tasks were created
+      const wasCreateCalled = await rig.waitForToolCall(
+        TRACKER_CREATE_TASK_TOOL_NAME,
+      );
+      expect(wasCreateCalled).toBe(true);
+
+      const toolLogs = rig.readToolLogs();
+      const createCalls = toolLogs.filter(
+        (l) => l.toolRequest.name === TRACKER_CREATE_TASK_TOOL_NAME,
+      );
+      expect(createCalls.length).toBeGreaterThanOrEqual(3);
+
+      await rig.expectToolCallSuccess([
+        TEST_AGENTS.I18N_AGENT.name,
+        TEST_AGENTS.SECURITY_AGENT.name,
+        TEST_AGENTS.CSS_AGENT.name,
+      ]);
+    },
+  });
+});

integration-tests/run_shell_command_file_stream.test.ts

@@ -0,0 +1,275 @@
+/**
+ * @license
+ * Copyright 2026 Google LLC
+ * SPDX-License-Identifier: Apache-2.0
+ */
+import { describe, it, expect, beforeEach, afterEach } from 'vitest';
+import { TestRig } from './test-helper.js';
+import * as fs from 'node:fs';
+import * as path from 'node:path';
+
+describe('run_shell_command streaming to file regression', () => {
+  let rig: TestRig;
+
+  beforeEach(() => {
+    rig = new TestRig();
+  });
+
+  afterEach(async () => await rig.cleanup());
+
+  it('should stream large outputs to a file and verify full content presence', async () => {
+    await rig.setup(
+      'should stream large outputs to a file and verify full content presence',
+      {
+        settings: { tools: { core: ['run_shell_command'] } },
+      },
+    );
+
+    const numLines = 20000;
+    const testFileName = 'large_output_test.txt';
+    const testFilePath = path.join(rig.testDir!, testFileName);
+
+    // Create a ~20MB file with unique content at start and end
+    const startMarker = 'START_OF_FILE_MARKER';
+    const endMarker = 'END_OF_FILE_MARKER';
+
+    const stream = fs.createWriteStream(testFilePath);
+    stream.write(startMarker + '\n');
+    for (let i = 0; i < numLines; i++) {
+      stream.write(`Line ${i + 1}: ` + 'A'.repeat(1000) + '\n');
+    }
+    stream.write(endMarker + '\n');
+    await new Promise((resolve) => stream.end(resolve));
+
+    const fileSize = fs.statSync(testFilePath).size;
+    expect(fileSize).toBeGreaterThan(20000000);
+
+    const prompt = `Use run_shell_command to cat ${testFileName} and say 'Done.'`;
+    await rig.run({ args: prompt });
+
+    let savedFilePath = '';
+    const tmpdir = path.join(rig.homeDir!, '.gemini', 'tmp');
+    if (fs.existsSync(tmpdir)) {
+      const findFiles = (dir: string): string[] => {
+        let results: string[] = [];
+        const list = fs.readdirSync(dir, { withFileTypes: true });
+        for (const file of list) {
+          const fullPath = path.join(dir, file.name);
+          if (file.isDirectory()) {
+            results = results.concat(findFiles(fullPath));
+          } else if (file.isFile() && file.name.endsWith('.txt')) {
+            results.push(fullPath);
+          } else if (file.isFile() && file.name.endsWith('.log')) {
+            results.push(fullPath);
+          }
+        }
+        return results;
+      };
+
+      const files = findFiles(tmpdir);
+      files.sort((a, b) => fs.statSync(b).mtimeMs - fs.statSync(a).mtimeMs);
+
+      for (const p of files) {
+        try {
+          const stat = fs.statSync(p);
+          if (stat.size >= 20000000) {
+            savedFilePath = p;
+            break;
+          }
+        } catch {
+          // ignore
+        }
+      }
+
+      if (!savedFilePath) {
+        const fileStats = files.map((p) => {
+          try {
+            return { p, size: fs.statSync(p).size };
+          } catch {
+            return { p, size: 'error' };
+          }
+        });
+        rig.log('Available files:', JSON.stringify(fileStats, null, 2));
+      }
+    }
+
+    expect(
+      savedFilePath,
+      `Expected to find a saved output file >= 20MB in ${tmpdir}`,
+    ).toBeTruthy();
+
+    const savedContent = fs.readFileSync(savedFilePath, 'utf8');
+    expect(savedContent).toContain(startMarker);
+    expect(savedContent).toContain(endMarker);
+    expect(savedContent.length).toBeGreaterThanOrEqual(fileSize);
+
+    fs.unlinkSync(savedFilePath);
+  }, 120000);
+
+  it('should stream very large (50MB) outputs to a file and verify full content presence', async () => {
+    await rig.setup(
+      'should stream very large (50MB) outputs to a file and verify full content presence',
+      {
+        settings: { tools: { core: ['run_shell_command'] } },
+      },
+    );
+
+    const numLines = 1000000;
+    const testFileName = 'very_large_output_test.txt';
+    const testFilePath = path.join(rig.testDir!, testFileName);
+
+    // Create a ~50MB file with unique content at start and end
+    const startMarker = 'START_OF_FILE_MARKER';
+    const endMarker = 'END_OF_FILE_MARKER';
+
+    const stream = fs.createWriteStream(testFilePath);
+    stream.write(startMarker + '\n');
+    for (let i = 0; i < numLines; i++) {
+      stream.write(`Line ${i + 1}: ` + 'A'.repeat(40) + '\n');
+    }
+    stream.write(endMarker + '\n');
+    await new Promise((resolve) => stream.end(resolve));
+
+    const fileSize = fs.statSync(testFilePath).size;
+    expect(fileSize).toBeGreaterThan(45000000);
+
+    const prompt = `Use run_shell_command to cat ${testFileName} and say 'Done.'`;
+    await rig.run({ args: prompt });
+
+    let savedFilePath = '';
+    const tmpdir = path.join(rig.homeDir!, '.gemini', 'tmp');
+    if (fs.existsSync(tmpdir)) {
+      const findFiles = (dir: string): string[] => {
+        let results: string[] = [];
+        const list = fs.readdirSync(dir, { withFileTypes: true });
+        for (const file of list) {
+          const fullPath = path.join(dir, file.name);
+          if (file.isDirectory()) {
+            results = results.concat(findFiles(fullPath));
+          } else if (file.isFile() && file.name.endsWith('.txt')) {
+            results.push(fullPath);
+          } else if (file.isFile() && file.name.endsWith('.log')) {
+            results.push(fullPath);
+          }
+        }
+        return results;
+      };
+
+      const files = findFiles(tmpdir);
+      files.sort((a, b) => fs.statSync(b).mtimeMs - fs.statSync(a).mtimeMs);
+
+      for (const p of files) {
+        try {
+          const stat = fs.statSync(p);
+          if (stat.size >= 20000000) {
+            savedFilePath = p;
+            break;
+          }
+        } catch {
+          // ignore
+        }
+      }
+
+      if (!savedFilePath) {
+        const fileStats = files.map((p) => {
+          try {
+            return { p, size: fs.statSync(p).size };
+          } catch {
+            return { p, size: 'error' };
+          }
+        });
+        rig.log('Available files:', JSON.stringify(fileStats, null, 2));
+      }
+    }
+
+    expect(
+      savedFilePath,
+      `Expected to find a saved output file >= 20MB in ${tmpdir}`,
+    ).toBeTruthy();
+
+    const savedContent = fs.readFileSync(savedFilePath, 'utf8');
+    expect(savedContent).toContain(startMarker);
+    expect(savedContent).toContain(endMarker);
+    expect(savedContent.length).toBeGreaterThanOrEqual(fileSize);
+
+    fs.unlinkSync(savedFilePath);
+  }, 120000);
+
+  it('should produce clean output resolving carriage returns and backspaces', async () => {
+    await rig.setup(
+      'should produce clean output resolving carriage returns and backspaces',
+      {
+        settings: {
+          tools: { core: ['run_shell_command'] },
+        },
+      },
+    );
+
+    const script = `
+import sys
+import time
+
+# Fill buffer to force file streaming/truncation
+# 45000 chars to be safe (default threshold is 40000)
+print('A' * 45000)
+sys.stdout.flush()
+
+# Test sequence
+print('XXXXX', end='', flush=True)
+time.sleep(0.5)
+print('\\rYYYYY', end='', flush=True)
+time.sleep(0.5)
+print('\\nNext Line', end='', flush=True)
+`;
+    const scriptPath = path.join(rig.testDir!, 'test_script.py');
+    fs.writeFileSync(scriptPath, script);
+
+    const prompt = `run_shell_command python3 "${scriptPath}"`;
+    await rig.run({ args: prompt });
+
+    let savedFilePath = '';
+    const tmpdir = path.join(rig.homeDir!, '.gemini', 'tmp');
+    if (fs.existsSync(tmpdir)) {
+      const findFiles = (dir: string): string[] => {
+        let results: string[] = [];
+        const list = fs.readdirSync(dir, { withFileTypes: true });
+        for (const file of list) {
+          const fullPath = path.join(dir, file.name);
+          if (file.isDirectory()) {
+            results = results.concat(findFiles(fullPath));
+          } else if (file.isFile() && file.name.endsWith('.txt')) {
+            results.push(fullPath);
+          }
+        }
+        return results;
+      };
+
+      const files = findFiles(tmpdir);
+      files.sort((a, b) => fs.statSync(b).mtimeMs - fs.statSync(a).mtimeMs);
+
+      if (files.length > 0) {
+        savedFilePath = files[0];
+      }
+    }
+
+    expect(savedFilePath, 'Output file should exist').toBeTruthy();
+    const content = fs.readFileSync(savedFilePath, 'utf8');
+
+    // Verify it contains the large chunk
+    expect(content).toContain('AAAA');
+
+    // Verify cleanup logic:
+    // 1. The final text "YYYYY" should be present.
+    expect(content).toContain('YYYYY');
+    // 2. The next line should be present.
+    expect(content).toContain('Next Line');
+
+    // 3. Verify overwrite happened.
+    // In raw output, we would have "XXXXX...YYYYY".
+    // In processed output, "YYYYY" overwrites "XXXXX".
+    // We confirm that escape codes are stripped (processed text).
+
+    // 4. Check for ANSI escape codes (like \\x1b) just in case
+    expect(content).not.toContain('\x1b');
+  }, 60000);
+});

package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "@google/gemini-cli",
-  "version": "0.39.0-nightly.20260408.e77b22e63",
+  "version": "0.40.0-nightly.20260414.g5b1f7375a",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "@google/gemini-cli",
-      "version": "0.39.0-nightly.20260408.e77b22e63",
+      "version": "0.40.0-nightly.20260414.g5b1f7375a",
       "workspaces": [
         "packages/*"
       ],
@@ -1727,29 +1727,6 @@
         "node": ">=8"
       }
     },
-    "node_modules/@joshua.litt/get-ripgrep": {
-      "version": "0.0.3",
-      "resolved": "https://registry.npmjs.org/@joshua.litt/get-ripgrep/-/get-ripgrep-0.0.3.tgz",
-      "integrity": "sha512-rycdieAKKqXi2bsM7G2ayDiNk5CAX8ZOzsTQsirfOqUKPef04Xw40BWGGyimaOOuvPgLWYt3tPnLLG3TvPXi5Q==",
-      "license": "MIT",
-      "dependencies": {
-        "@lvce-editor/verror": "^1.6.0",
-        "execa": "^9.5.2",
-        "extract-zip": "^2.0.1",
-        "fs-extra": "^11.3.0",
-        "got": "^14.4.5",
-        "path-exists": "^5.0.0",
-        "xdg-basedir": "^5.1.0"
-      }
-    },
-    "node_modules/@joshua.litt/get-ripgrep/node_modules/path-exists": {
-      "version": "5.0.0",
-      "resolved": "https://registry.npmjs.org/path-exists/-/path-exists-5.0.0.tgz",
-      "integrity": "sha512-RjhtfwJOxzcFmNOi6ltcbcu4Iu+FL3zEj83dk4kAS+fVpTxXLO1b38RvJgT/0QwvV/L3aY9TAnyv0EOqW4GoMQ==",
-      "engines": {
-        "node": "^12.20.0 || ^14.13.1 || >=16.0.0"
-      }
-    },
     "node_modules/@jridgewell/gen-mapping": {
       "version": "0.3.13",
       "resolved": "https://registry.npmjs.org/@jridgewell/gen-mapping/-/gen-mapping-0.3.13.tgz",
@@ -1932,12 +1909,6 @@
       "integrity": "sha512-GaHYm+c0O9MjZRu0ongGBRbinu8gVAMd2UZjji6jVmqKtZluZnptXGWhz1E8j8D2HJ3f/yMxKAUC0b+57wncIw==",
       "license": "MIT"
     },
-    "node_modules/@lvce-editor/verror": {
-      "version": "1.7.0",
-      "resolved": "https://registry.npmjs.org/@lvce-editor/verror/-/verror-1.7.0.tgz",
-      "integrity": "sha512-+LGuAEIC2L7pbvkyAQVWM2Go0dAy+UWEui28g07zNtZsCBhm+gusBK8PNwLJLV5Jay+TyUYuwLIbJdjLLzqEBg==",
-      "license": "MIT"
-    },
     "node_modules/@lydell/node-pty": {
       "version": "1.1.0",
       "resolved": "https://registry.npmjs.org/@lydell/node-pty/-/node-pty-1.1.0.tgz",
@@ -3590,18 +3561,6 @@
         "url": "https://ko-fi.com/killymxi"
       }
     },
-    "node_modules/@sindresorhus/is": {
-      "version": "7.0.2",
-      "resolved": "https://registry.npmjs.org/@sindresorhus/is/-/is-7.0.2.tgz",
-      "integrity": "sha512-d9xRovfKNz1SKieM0qJdO+PQonjnnIfSNWfHYnBSJ9hkjm0ZPw6HlxscDXYstp3z+7V2GOFHc+J0CYrYTjqCJw==",
-      "license": "MIT",
-      "engines": {
-        "node": ">=18"
-      },
-      "funding": {
-        "url": "https://github.com/sindresorhus/is?sponsor=1"
-      }
-    },
     "node_modules/@sindresorhus/merge-streams": {
       "version": "4.0.0",
       "resolved": "https://registry.npmjs.org/@sindresorhus/merge-streams/-/merge-streams-4.0.0.tgz",
@@ -3614,18 +3573,6 @@
         "url": "https://github.com/sponsors/sindresorhus"
       }
     },
-    "node_modules/@szmarczak/http-timer": {
-      "version": "5.0.1",
-      "resolved": "https://registry.npmjs.org/@szmarczak/http-timer/-/http-timer-5.0.1.tgz",
-      "integrity": "sha512-+PmQX0PiAYPMeVYe237LJAYvOMYW1j2rH5YROyS3b4CTVJum34HfRvKvAzozHAQG0TnHNdUfY9nCeUyRAs//cw==",
-      "license": "MIT",
-      "dependencies": {
-        "defer-to-connect": "^2.0.1"
-      },
-      "engines": {
-        "node": ">=14.16"
-      }
-    },
     "node_modules/@textlint/ast-node-types": {
       "version": "15.2.2",
       "resolved": "https://registry.npmjs.org/@textlint/ast-node-types/-/ast-node-types-15.2.2.tgz",
@@ -3931,12 +3878,6 @@
       "integrity": "sha512-pUY3cKH/Nm2yYrEmDlPR1mR7yszjGx4DrwPjQ702C4/D5CwHuZTgZdIdwPkRbcuhs7BAh2L5rg3CL5cbRiGTCQ==",
       "license": "MIT"
     },
-    "node_modules/@types/http-cache-semantics": {
-      "version": "4.0.4",
-      "resolved": "https://registry.npmjs.org/@types/http-cache-semantics/-/http-cache-semantics-4.0.4.tgz",
-      "integrity": "sha512-1m0bIFVc7eJWyve9S0RnuRgcQqF/Xd5QsUZAZeQFr1Q3/p9JWoQQEqmVy+DPTNpGXwhgIetAoYF8JSc33q29QA==",
-      "license": "MIT"
-    },
     "node_modules/@types/http-errors": {
       "version": "2.0.5",
       "resolved": "https://registry.npmjs.org/@types/http-errors/-/http-errors-2.0.5.tgz",
@@ -6008,33 +5949,6 @@
         "node": ">=8"
       }
     },
-    "node_modules/cacheable-lookup": {
-      "version": "7.0.0",
-      "resolved": "https://registry.npmjs.org/cacheable-lookup/-/cacheable-lookup-7.0.0.tgz",
-      "integrity": "sha512-+qJyx4xiKra8mZrcwhjMRMUhD5NR1R8esPkzIYxX96JiecFoxAXFuz/GpR3+ev4PE1WamHip78wV0vcmPQtp8w==",
-      "license": "MIT",
-      "engines": {
-        "node": ">=14.16"
-      }
-    },
-    "node_modules/cacheable-request": {
-      "version": "12.0.1",
-      "resolved": "https://registry.npmjs.org/cacheable-request/-/cacheable-request-12.0.1.tgz",
-      "integrity": "sha512-Yo9wGIQUaAfIbk+qY0X4cDQgCosecfBe3V9NSyeY4qPC2SAkbCS4Xj79VP8WOzitpJUZKc/wsRCYF5ariDIwkg==",
-      "license": "MIT",
-      "dependencies": {
-        "@types/http-cache-semantics": "^4.0.4",
-        "get-stream": "^9.0.1",
-        "http-cache-semantics": "^4.1.1",
-        "keyv": "^4.5.4",
-        "mimic-response": "^4.0.0",
-        "normalize-url": "^8.0.1",
-        "responselike": "^3.0.0"
-      },
-      "engines": {
-        "node": ">=18"
-      }
-    },
     "node_modules/call-bind": {
       "version": "1.0.8",
       "resolved": "https://registry.npmjs.org/call-bind/-/call-bind-1.0.8.tgz",
@@ -6968,33 +6882,6 @@
         }
       }
     },
-    "node_modules/decompress-response": {
-      "version": "6.0.0",
-      "resolved": "https://registry.npmjs.org/decompress-response/-/decompress-response-6.0.0.tgz",
-      "integrity": "sha512-aW35yZM6Bb/4oJlZncMH2LCoZtJXTRxES17vE3hoRiowU2kWHaJKFkSBDnDR+cm9J+9QhXmREyIfv0pji9ejCQ==",
-      "license": "MIT",
-      "dependencies": {
-        "mimic-response": "^3.1.0"
-      },
-      "engines": {
-        "node": ">=10"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/sindresorhus"
-      }
-    },
-    "node_modules/decompress-response/node_modules/mimic-response": {
-      "version": "3.1.0",
-      "resolved": "https://registry.npmjs.org/mimic-response/-/mimic-response-3.1.0.tgz",
-      "integrity": "sha512-z0yWI+4FDrrweS8Zmt4Ej5HdJmky15+L2e6Wgn3+iK5fWzb6T3fhNFq2+MeTRb064c6Wr4N/wv0DzQTjNzHNGQ==",
-      "license": "MIT",
-      "engines": {
-        "node": ">=10"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/sindresorhus"
-      }
-    },
     "node_modules/deep-eql": {
       "version": "5.0.2",
       "resolved": "https://registry.npmjs.org/deep-eql/-/deep-eql-5.0.2.tgz",
@@ -7057,15 +6944,6 @@
         "url": "https://github.com/sponsors/sindresorhus"
       }
     },
-    "node_modules/defer-to-connect": {
-      "version": "2.0.1",
-      "resolved": "https://registry.npmjs.org/defer-to-connect/-/defer-to-connect-2.0.1.tgz",
-      "integrity": "sha512-4tvttepXG1VaYGrRibk5EwJd1t4udunSOVMdLSAL6mId1ix438oPwPZMALY41FCijukO1L0twNcGsdzS7dHgDg==",
-      "license": "MIT",
-      "engines": {
-        "node": ">=10"
-      }
-    },
     "node_modules/define-data-property": {
       "version": "1.1.4",
       "resolved": "https://registry.npmjs.org/define-data-property/-/define-data-property-1.1.4.tgz",
@@ -8428,9 +8306,9 @@
       }
     },
     "node_modules/execa": {
-      "version": "9.6.0",
-      "resolved": "https://registry.npmjs.org/execa/-/execa-9.6.0.tgz",
-      "integrity": "sha512-jpWzZ1ZhwUmeWRhS7Qv3mhpOhLfwI+uAX4e5fOcXqwMR7EcJ0pj2kV1CVzHVMX/LphnKWD3LObjZCoJ71lKpHw==",
+      "version": "9.6.1",
+      "resolved": "https://registry.npmjs.org/execa/-/execa-9.6.1.tgz",
+      "integrity": "sha512-9Be3ZoN4LmYR90tUoVu2te2BsbzHfhJyfEiAVfz7N5/zv+jduIfLrV2xdQXOHbaD6KgpGdO9PRPM1Y4Q9QkPkA==",
       "license": "MIT",
       "dependencies": {
         "@sindresorhus/merge-streams": "^4.0.0",
@@ -8950,15 +8828,6 @@
         "node": ">= 6"
       }
     },
-    "node_modules/form-data-encoder": {
-      "version": "4.1.0",
-      "resolved": "https://registry.npmjs.org/form-data-encoder/-/form-data-encoder-4.1.0.tgz",
-      "integrity": "sha512-G6NsmEW15s0Uw9XnCg+33H3ViYRyiM0hMrMhhqQOR8NFc5GhYrI+6I3u7OTw7b91J2g8rtvMBZJDbcGb2YUniw==",
-      "license": "MIT",
-      "engines": {
-        "node": ">= 18"
-      }
-    },
     "node_modules/form-data/node_modules/mime-types": {
       "version": "2.1.35",
       "resolved": "https://registry.npmjs.org/mime-types/-/mime-types-2.1.35.tgz",
@@ -9586,43 +9455,6 @@
         "url": "https://github.com/sponsors/ljharb"
       }
     },
-    "node_modules/got": {
-      "version": "14.4.8",
-      "resolved": "https://registry.npmjs.org/got/-/got-14.4.8.tgz",
-      "integrity": "sha512-vxwU4HuR0BIl+zcT1LYrgBjM+IJjNElOjCzs0aPgHorQyr/V6H6Y73Sn3r3FOlUffvWD+Q5jtRuGWaXkU8Jbhg==",
-      "license": "MIT",
-      "dependencies": {
-        "@sindresorhus/is": "^7.0.1",
-        "@szmarczak/http-timer": "^5.0.1",
-        "cacheable-lookup": "^7.0.0",
-        "cacheable-request": "^12.0.1",
-        "decompress-response": "^6.0.0",
-        "form-data-encoder": "^4.0.2",
-        "http2-wrapper": "^2.2.1",
-        "lowercase-keys": "^3.0.0",
-        "p-cancelable": "^4.0.1",
-        "responselike": "^3.0.0",
-        "type-fest": "^4.26.1"
-      },
-      "engines": {
-        "node": ">=20"
-      },
-      "funding": {
-        "url": "https://github.com/sindresorhus/got?sponsor=1"
-      }
-    },
-    "node_modules/got/node_modules/type-fest": {
-      "version": "4.41.0",
-      "resolved": "https://registry.npmjs.org/type-fest/-/type-fest-4.41.0.tgz",
-      "integrity": "sha512-TeTSQ6H5YHvpqVwBRcnLDCBnDOHWYu7IvGbHT6N8AOymcr9PJGjc1GTtiWZTYg0NCgYwvnYWEkVChQAr9bjfwA==",
-      "license": "(MIT OR CC0-1.0)",
-      "engines": {
-        "node": ">=16"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/sindresorhus"
-      }
-    },
     "node_modules/graceful-fs": {
       "version": "4.2.11",
       "resolved": "https://registry.npmjs.org/graceful-fs/-/graceful-fs-4.2.11.tgz",
@@ -9878,12 +9710,6 @@
         "entities": "^4.4.0"
       }
     },
-    "node_modules/http-cache-semantics": {
-      "version": "4.2.0",
-      "resolved": "https://registry.npmjs.org/http-cache-semantics/-/http-cache-semantics-4.2.0.tgz",
-      "integrity": "sha512-dTxcvPXqPvXBQpq5dUr6mEMJX4oIEFv6bwom3FDwKRDsuIjjJGANqhBuoAn9c1RQJIdAKav33ED65E2ys+87QQ==",
-      "license": "BSD-2-Clause"
-    },
     "node_modules/http-errors": {
       "version": "2.0.1",
       "resolved": "https://registry.npmjs.org/http-errors/-/http-errors-2.0.1.tgz",
@@ -9917,19 +9743,6 @@
         "node": ">= 14"
       }
     },
-    "node_modules/http2-wrapper": {
-      "version": "2.2.1",
-      "resolved": "https://registry.npmjs.org/http2-wrapper/-/http2-wrapper-2.2.1.tgz",
-      "integrity": "sha512-V5nVw1PAOgfI3Lmeaj2Exmeg7fenjhRUgz1lPSezy1CuhPYbgQtbQj4jZfEAEMlaL+vupsvhjqCyjzob0yxsmQ==",
-      "license": "MIT",
-      "dependencies": {
-        "quick-lru": "^5.1.1",
-        "resolve-alpn": "^1.2.0"
-      },
-      "engines": {
-        "node": ">=10.19.0"
-      }
-    },
     "node_modules/https-proxy-agent": {
       "version": "7.0.6",
       "resolved": "https://registry.npmjs.org/https-proxy-agent/-/https-proxy-agent-7.0.6.tgz",
@@ -11040,6 +10853,7 @@
       "version": "3.0.1",
       "resolved": "https://registry.npmjs.org/json-buffer/-/json-buffer-3.0.1.tgz",
       "integrity": "sha512-4bV5BfR2mqfQTJm+V5tPPdf+ZpuhiIvTuAB5g8kcrXOZpTT/QwwVRWBywX1ozr6lEuPdbHxwaJlm9G6mI2sfSQ==",
+      "dev": true,
       "license": "MIT"
     },
     "node_modules/json-parse-better-errors": {
@@ -11235,6 +11049,7 @@
       "version": "4.5.4",
       "resolved": "https://registry.npmjs.org/keyv/-/keyv-4.5.4.tgz",
       "integrity": "sha512-oxVHkHR/EJf2CNXnWxRLW6mg7JyCCUcG0DtEGmL2ctUo1PNTin1PUil+r/+4r5MpVgC/fn1kjsx7mjSujKqIpw==",
+      "dev": true,
       "license": "MIT",
       "dependencies": {
         "json-buffer": "3.0.1"
@@ -11699,18 +11514,6 @@
       "integrity": "sha512-CdzqowRJCeLU72bHvWqwRBBlLcMEtIvGrlvef74kMnV2AolS9Y8xUv1I0U/MNAWMhBlKIoyuEgoJ0t/bbwHbLQ==",
       "license": "MIT"
     },
-    "node_modules/lowercase-keys": {
-      "version": "3.0.0",
-      "resolved": "https://registry.npmjs.org/lowercase-keys/-/lowercase-keys-3.0.0.tgz",
-      "integrity": "sha512-ozCC6gdQ+glXOQsveKD0YsDy8DSQFjDTz4zyzEHNV5+JP5D62LmfDZ6o1cycFx9ouG940M5dE8C8CTewdj2YWQ==",
-      "license": "MIT",
-      "engines": {
-        "node": "^12.20.0 || ^14.13.1 || >=16.0.0"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/sindresorhus"
-      }
-    },
     "node_modules/lowlight": {
       "version": "3.3.0",
       "resolved": "https://registry.npmjs.org/lowlight/-/lowlight-3.3.0.tgz",
@@ -11981,18 +11784,6 @@
         "url": "https://github.com/sponsors/sindresorhus"
       }
     },
-    "node_modules/mimic-response": {
-      "version": "4.0.0",
-      "resolved": "https://registry.npmjs.org/mimic-response/-/mimic-response-4.0.0.tgz",
-      "integrity": "sha512-e5ISH9xMYU0DzrT+jl8q2ze9D6eWBto+I8CNpe+VI+K2J/F/k3PdkdTdz4wvGVH4NTpo+NRYTVIuMQEMMcsLqg==",
-      "license": "MIT",
-      "engines": {
-        "node": "^12.20.0 || ^14.13.1 || >=16.0.0"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/sindresorhus"
-      }
-    },
     "node_modules/minimatch": {
       "version": "10.2.4",
       "resolved": "https://registry.npmjs.org/minimatch/-/minimatch-10.2.4.tgz",
@@ -12371,18 +12162,6 @@
         "node": "^16.14.0 || >=18.0.0"
       }
     },
-    "node_modules/normalize-url": {
-      "version": "8.0.2",
-      "resolved": "https://registry.npmjs.org/normalize-url/-/normalize-url-8.0.2.tgz",
-      "integrity": "sha512-Ee/R3SyN4BuynXcnTaekmaVdbDAEiNrHqjQIA37mHU8G9pf7aaAD4ZX3XjBLo6rsdcxA/gtkcNYZLt30ACgynw==",
-      "license": "MIT",
-      "engines": {
-        "node": ">=14.16"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/sindresorhus"
-      }
-    },
     "node_modules/npm-normalize-package-bin": {
       "version": "4.0.0",
       "resolved": "https://registry.npmjs.org/npm-normalize-package-bin/-/npm-normalize-package-bin-4.0.0.tgz",
@@ -12895,15 +12674,6 @@
         "url": "https://github.com/sponsors/ljharb"
       }
     },
-    "node_modules/p-cancelable": {
-      "version": "4.0.1",
-      "resolved": "https://registry.npmjs.org/p-cancelable/-/p-cancelable-4.0.1.tgz",
-      "integrity": "sha512-wBowNApzd45EIKdO1LaU+LrMBwAcjfPaYtVzV3lmfM3gf8Z4CHZsiIqlM8TZZ8okYvh5A1cP6gTfCRQtwUpaUg==",
-      "license": "MIT",
-      "engines": {
-        "node": ">=14.16"
-      }
-    },
     "node_modules/p-limit": {
       "version": "3.1.0",
       "resolved": "https://registry.npmjs.org/p-limit/-/p-limit-3.1.0.tgz",
@@ -13743,18 +13513,6 @@
       ],
       "license": "MIT"
     },
-    "node_modules/quick-lru": {
-      "version": "5.1.1",
-      "resolved": "https://registry.npmjs.org/quick-lru/-/quick-lru-5.1.1.tgz",
-      "integrity": "sha512-WuyALRjWPDGtt/wzJiadO5AXY+8hZ80hVpe6MyivgraREW751X3SbhRvG3eLKOYN+8VEvqLcf3wdnt44Z4S4SA==",
-      "license": "MIT",
-      "engines": {
-        "node": ">=10"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/sindresorhus"
-      }
-    },
     "node_modules/range-parser": {
       "version": "1.2.1",
       "resolved": "https://registry.npmjs.org/range-parser/-/range-parser-1.2.1.tgz",
@@ -14195,12 +13953,6 @@
         "url": "https://github.com/sponsors/ljharb"
       }
     },
-    "node_modules/resolve-alpn": {
-      "version": "1.2.1",
-      "resolved": "https://registry.npmjs.org/resolve-alpn/-/resolve-alpn-1.2.1.tgz",
-      "integrity": "sha512-0a1F4l73/ZFZOakJnQ3FvkJ2+gSTQWz/r2KE5OdDY0TxPm5h4GkqkWWfM47T7HsbnOtcJVEF4epCVy6u7Q3K+g==",
-      "license": "MIT"
-    },
     "node_modules/resolve-dir": {
       "version": "1.0.1",
       "resolved": "https://registry.npmjs.org/resolve-dir/-/resolve-dir-1.0.1.tgz",
@@ -14235,21 +13987,6 @@
         "url": "https://github.com/privatenumber/resolve-pkg-maps?sponsor=1"
       }
     },
-    "node_modules/responselike": {
-      "version": "3.0.0",
-      "resolved": "https://registry.npmjs.org/responselike/-/responselike-3.0.0.tgz",
-      "integrity": "sha512-40yHxbNcl2+rzXvZuVkrYohathsSJlMTXKryG5y8uciHv1+xDLHQpgjG64JUO9nrEq2jGLH6IZ8BcZyw3wrweg==",
-      "license": "MIT",
-      "dependencies": {
-        "lowercase-keys": "^3.0.0"
-      },
-      "engines": {
-        "node": ">=14.16"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/sindresorhus"
-      }
-    },
     "node_modules/restore-cursor": {
       "version": "4.0.0",
       "resolved": "https://registry.npmjs.org/restore-cursor/-/restore-cursor-4.0.0.tgz",
@@ -17644,18 +17381,6 @@
         }
       }
     },
-    "node_modules/xdg-basedir": {
-      "version": "5.1.0",
-      "resolved": "https://registry.npmjs.org/xdg-basedir/-/xdg-basedir-5.1.0.tgz",
-      "integrity": "sha512-GCPAHLvrIH13+c0SuacwvRYj2SxJXQ4kaVTT5xgL3kPrz56XxkF21IGhjSE1+W0aw7gpBWRGXLCPnPby6lSpmQ==",
-      "license": "MIT",
-      "engines": {
-        "node": ">=12"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/sindresorhus"
-      }
-    },
     "node_modules/xml2js": {
       "version": "0.5.0",
       "resolved": "https://registry.npmjs.org/xml2js/-/xml2js-0.5.0.tgz",
@@ -17864,7 +17589,7 @@
     },
     "packages/a2a-server": {
       "name": "@google/gemini-cli-a2a-server",
-      "version": "0.39.0-nightly.20260408.e77b22e63",
+      "version": "0.40.0-nightly.20260414.g5b1f7375a",
       "dependencies": {
         "@a2a-js/sdk": "0.3.11",
         "@google-cloud/storage": "^7.16.0",
@@ -17979,7 +17704,7 @@
     },
     "packages/cli": {
       "name": "@google/gemini-cli",
-      "version": "0.39.0-nightly.20260408.e77b22e63",
+      "version": "0.40.0-nightly.20260414.g5b1f7375a",
       "license": "Apache-2.0",
       "dependencies": {
         "@agentclientprotocol/sdk": "^0.16.1",
@@ -18079,44 +17804,6 @@
         "url": "https://github.com/sponsors/sindresorhus"
       }
     },
-    "packages/cli/node_modules/execa": {
-      "version": "9.6.1",
-      "resolved": "https://registry.npmjs.org/execa/-/execa-9.6.1.tgz",
-      "integrity": "sha512-9Be3ZoN4LmYR90tUoVu2te2BsbzHfhJyfEiAVfz7N5/zv+jduIfLrV2xdQXOHbaD6KgpGdO9PRPM1Y4Q9QkPkA==",
-      "license": "MIT",
-      "dependencies": {
-        "@sindresorhus/merge-streams": "^4.0.0",
-        "cross-spawn": "^7.0.6",
-        "figures": "^6.1.0",
-        "get-stream": "^9.0.0",
-        "human-signals": "^8.0.1",
-        "is-plain-obj": "^4.1.0",
-        "is-stream": "^4.0.1",
-        "npm-run-path": "^6.0.0",
-        "pretty-ms": "^9.2.0",
-        "signal-exit": "^4.1.0",
-        "strip-final-newline": "^4.0.0",
-        "yoctocolors": "^2.1.1"
-      },
-      "engines": {
-        "node": "^18.19.0 || >=20.5.0"
-      },
-      "funding": {
-        "url": "https://github.com/sindresorhus/execa?sponsor=1"
-      }
-    },
-    "packages/cli/node_modules/is-stream": {
-      "version": "4.0.1",
-      "resolved": "https://registry.npmjs.org/is-stream/-/is-stream-4.0.1.tgz",
-      "integrity": "sha512-Dnz92NInDqYckGEUJv689RbRiTSEHCQ7wOVeALbkOz999YpqT46yMRIGtSNl2iCL1waAZSx40+h59NV/EwzV/A==",
-      "license": "MIT",
-      "engines": {
-        "node": ">=18"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/sindresorhus"
-      }
-    },
     "packages/cli/node_modules/string-width": {
       "version": "8.1.0",
       "resolved": "https://registry.npmjs.org/string-width/-/string-width-8.1.0.tgz",
@@ -18151,7 +17838,7 @@
     },
     "packages/core": {
       "name": "@google/gemini-cli-core",
-      "version": "0.39.0-nightly.20260408.e77b22e63",
+      "version": "0.40.0-nightly.20260414.g5b1f7375a",
       "license": "Apache-2.0",
       "dependencies": {
         "@a2a-js/sdk": "0.3.11",
@@ -18162,7 +17849,6 @@
         "@google/genai": "1.30.0",
         "@grpc/grpc-js": "^1.14.3",
         "@iarna/toml": "^2.2.5",
-        "@joshua.litt/get-ripgrep": "^0.0.3",
         "@modelcontextprotocol/sdk": "^1.23.0",
         "@opentelemetry/api": "^1.9.0",
         "@opentelemetry/api-logs": "^0.211.0",
@@ -18190,6 +17876,7 @@
         "diff": "^8.0.3",
         "dotenv": "^17.2.4",
         "dotenv-expand": "^12.0.3",
+        "execa": "^9.6.1",
         "fast-levenshtein": "^2.0.6",
         "fdir": "^6.4.6",
         "fzf": "^0.5.2",
@@ -18418,7 +18105,7 @@
     },
     "packages/devtools": {
       "name": "@google/gemini-cli-devtools",
-      "version": "0.39.0-nightly.20260408.e77b22e63",
+      "version": "0.40.0-nightly.20260414.g5b1f7375a",
       "license": "Apache-2.0",
       "dependencies": {
         "ws": "^8.16.0"
@@ -18433,7 +18120,7 @@
     },
     "packages/sdk": {
       "name": "@google/gemini-cli-sdk",
-      "version": "0.39.0-nightly.20260408.e77b22e63",
+      "version": "0.40.0-nightly.20260414.g5b1f7375a",
       "license": "Apache-2.0",
       "dependencies": {
         "@google/gemini-cli-core": "file:../core",
@@ -18450,7 +18137,7 @@
     },
     "packages/test-utils": {
       "name": "@google/gemini-cli-test-utils",
-      "version": "0.39.0-nightly.20260408.e77b22e63",
+      "version": "0.40.0-nightly.20260414.g5b1f7375a",
       "license": "Apache-2.0",
       "dependencies": {
         "@google/gemini-cli-core": "file:../core",
@@ -18468,7 +18155,7 @@
     },
     "packages/vscode-ide-companion": {
       "name": "gemini-cli-vscode-ide-companion",
-      "version": "0.39.0-nightly.20260408.e77b22e63",
+      "version": "0.40.0-nightly.20260414.g5b1f7375a",
       "license": "LICENSE",
       "dependencies": {
         "@modelcontextprotocol/sdk": "^1.23.0",

package.json

@@ -1,6 +1,6 @@
 {
   "name": "@google/gemini-cli",
-  "version": "0.39.0-nightly.20260408.e77b22e63",
+  "version": "0.40.0-nightly.20260414.g5b1f7375a",
   "engines": {
     "node": ">=20.0.0"
   },
@@ -14,7 +14,7 @@
     "url": "git+https://github.com/google-gemini/gemini-cli.git"
   },
   "config": {
-    "sandboxImageUri": "us-docker.pkg.dev/gemini-code-dev/gemini-cli/sandbox:0.39.0-nightly.20260408.e77b22e63"
+    "sandboxImageUri": "us-docker.pkg.dev/gemini-code-dev/gemini-cli/sandbox:0.40.0-nightly.20260414.g5b1f7375a"
   },
   "scripts": {
     "start": "cross-env NODE_ENV=development node scripts/start.js",

packages/a2a-server/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@google/gemini-cli-a2a-server",
-  "version": "0.39.0-nightly.20260408.e77b22e63",
+  "version": "0.40.0-nightly.20260414.g5b1f7375a",
   "description": "Gemini CLI A2A Server",
   "repository": {
     "type": "git",

packages/cli/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@google/gemini-cli",
-  "version": "0.39.0-nightly.20260408.e77b22e63",
+  "version": "0.40.0-nightly.20260414.g5b1f7375a",
   "description": "Gemini CLI",
   "license": "Apache-2.0",
   "repository": {
@@ -27,7 +27,7 @@
     "dist"
   ],
   "config": {
-    "sandboxImageUri": "us-docker.pkg.dev/gemini-code-dev/gemini-cli/sandbox:0.39.0-nightly.20260408.e77b22e63"
+    "sandboxImageUri": "us-docker.pkg.dev/gemini-code-dev/gemini-cli/sandbox:0.40.0-nightly.20260414.g5b1f7375a"
   },
   "dependencies": {
     "@agentclientprotocol/sdk": "^0.16.1",

packages/cli/src/ui/auth/ApiAuthDialog.test.tsx

@@ -4,7 +4,7 @@
  * SPDX-License-Identifier: Apache-2.0
  */
 
-import { render } from '../../test-utils/render.js';
+import { renderWithProviders } from '../../test-utils/render.js';
 import { waitFor } from '../../test-utils/async.js';
 import { describe, it, expect, vi, beforeEach, type Mock } from 'vitest';
 import { ApiAuthDialog } from './ApiAuthDialog.js';
@@ -40,11 +40,16 @@ vi.mock('../components/shared/text-buffer.js', async (importOriginal) => {
   };
 });
 
-vi.mock('../contexts/UIStateContext.js', () => ({
-  useUIState: vi.fn(() => ({
-    terminalWidth: 80,
-  })),
-}));
+vi.mock('../contexts/UIStateContext.js', async (importOriginal) => {
+  const actual =
+    await importOriginal<typeof import('../contexts/UIStateContext.js')>();
+  return {
+    ...actual,
+    useUIState: vi.fn(() => ({
+      terminalWidth: 80,
+    })),
+  };
+});
 
 const mockedUseKeypress = useKeypress as Mock;
 const mockedUseTextBuffer = useTextBuffer as Mock;
@@ -73,7 +78,7 @@ describe('ApiAuthDialog', () => {
   });
 
   it('renders correctly', async () => {
-    const { lastFrame, unmount } = await render(
+    const { lastFrame, unmount } = await renderWithProviders(
       <ApiAuthDialog onSubmit={onSubmit} onCancel={onCancel} />,
     );
     expect(lastFrame()).toMatchSnapshot();
@@ -81,7 +86,7 @@ describe('ApiAuthDialog', () => {
   });
 
   it('renders with a defaultValue', async () => {
-    const { unmount } = await render(
+    const { unmount } = await renderWithProviders(
       <ApiAuthDialog
         onSubmit={onSubmit}
         onCancel={onCancel}
@@ -111,7 +116,7 @@ describe('ApiAuthDialog', () => {
     'calls $expectedCall.name when $keyName is pressed',
     async ({ keyName, sequence, expectedCall, args }) => {
       mockBuffer.text = 'submitted-key'; // Set for the onSubmit case
-      const { unmount } = await render(
+      const { unmount } = await renderWithProviders(
         <ApiAuthDialog onSubmit={onSubmit} onCancel={onCancel} />,
       );
       // calls[0] is the ApiAuthDialog's useKeypress (Ctrl+C handler)
@@ -133,7 +138,7 @@ describe('ApiAuthDialog', () => {
   );
 
   it('displays an error message', async () => {
-    const { lastFrame, unmount } = await render(
+    const { lastFrame, unmount } = await renderWithProviders(
       <ApiAuthDialog
         onSubmit={onSubmit}
         onCancel={onCancel}
@@ -146,7 +151,7 @@ describe('ApiAuthDialog', () => {
   });
 
   it('calls clearApiKey and clears buffer when Ctrl+C is pressed', async () => {
-    const { unmount } = await render(
+    const { unmount } = await renderWithProviders(
       <ApiAuthDialog onSubmit={onSubmit} onCancel={onCancel} />,
     );
     // Call 0 is ApiAuthDialog (isActive: true)

packages/cli/src/ui/components/AnsiOutput.tsx

@@ -4,7 +4,7 @@
  * SPDX-License-Identifier: Apache-2.0
  */
 
-import type React from 'react';
+import React from 'react';
 import { Box, Text } from 'ink';
 import type { AnsiLine, AnsiOutput, AnsiToken } from '@google/gemini-cli-core';
 
@@ -53,23 +53,26 @@ export const AnsiOutputText: React.FC<AnsiOutputProps> = ({
   );
 };
 
-export const AnsiLineText: React.FC<{ line: AnsiLine }> = ({ line }) => (
-  <Text>
-    {line.length > 0
-      ? line.map((token: AnsiToken, tokenIndex: number) => (
-          <Text
-            key={tokenIndex}
-            color={token.fg}
-            backgroundColor={token.bg}
-            inverse={token.inverse}
-            dimColor={token.dim}
-            bold={token.bold}
-            italic={token.italic}
-            underline={token.underline}
-          >
-            {token.text}
-          </Text>
-        ))
-      : null}
-  </Text>
+export const AnsiLineText = React.memo<{ line: AnsiLine }>(
+  ({ line }: { line: AnsiLine }) => (
+    <Text>
+      {line.length > 0
+        ? line.map((token: AnsiToken, tokenIndex: number) => (
+            <Text
+              key={tokenIndex}
+              color={token.fg}
+              backgroundColor={token.bg}
+              inverse={token.inverse}
+              dimColor={token.dim}
+              bold={token.bold}
+              italic={token.italic}
+              underline={token.underline}
+            >
+              {token.text}
+            </Text>
+          ))
+        : null}
+    </Text>
+  ),
 );
+AnsiLineText.displayName = 'AnsiLineText';

packages/cli/src/ui/components/AskUserDialog.tsx

@@ -13,7 +13,8 @@ import {
   useReducer,
   useContext,
 } from 'react';
-import { Box, Text } from 'ink';
+import { Box, Text, type DOMElement } from 'ink';
+import { useMouseClick } from '../hooks/useMouseClick.js';
 import { theme } from '../semantic-colors.js';
 import { checkExhaustive, type Question } from '@google/gemini-cli-core';
 import { BaseSelectionList } from './shared/BaseSelectionList.js';
@@ -85,6 +86,24 @@ function autoBoldIfPlain(text: string): string {
   return text;
 }
 
+const ClickableCheckbox: React.FC<{
+  isChecked: boolean;
+  onClick: () => void;
+}> = ({ isChecked, onClick }) => {
+  const ref = useRef<DOMElement>(null);
+  useMouseClick(ref, () => {
+    onClick();
+  });
+
+  return (
+    <Box ref={ref}>
+      <Text color={isChecked ? theme.status.success : theme.text.secondary}>
+        [{isChecked ? 'x' : ' '}]
+      </Text>
+    </Box>
+  );
+};
+
 interface AskUserDialogState {
   answers: { [key: string]: string };
   isEditingCustomOption: boolean;
@@ -919,13 +938,14 @@ const ChoiceQuestionView: React.FC<ChoiceQuestionViewProps> = ({
             return (
               <Box flexDirection="row">
                 {showCheck && (
-                  <Text
-                    color={
-                      isChecked ? theme.status.success : theme.text.secondary
-                    }
-                  >
-                    [{isChecked ? 'x' : ' '}]
-                  </Text>
+                  <ClickableCheckbox
+                    isChecked={isChecked}
+                    onClick={() => {
+                      if (!context.isSelected) {
+                        handleSelect(optionItem);
+                      }
+                    }}
+                  />
                 )}
                 <Text color={theme.text.primary}> </Text>
                 <TextInput
@@ -966,13 +986,14 @@ const ChoiceQuestionView: React.FC<ChoiceQuestionViewProps> = ({
             <Box flexDirection="column">
               <Box flexDirection="row">
                 {showCheck && (
-                  <Text
-                    color={
-                      isChecked ? theme.status.success : theme.text.secondary
-                    }
-                  >
-                    [{isChecked ? 'x' : ' '}]
-                  </Text>
+                  <ClickableCheckbox
+                    isChecked={isChecked}
+                    onClick={() => {
+                      if (!context.isSelected) {
+                        handleSelect(optionItem);
+                      }
+                    }}
+                  />
                 )}
                 <Text color={labelColor} bold={optionItem.type === 'done'}>
                   {' '}

packages/cli/src/ui/components/InputPrompt.test.tsx

@@ -3877,8 +3877,9 @@ describe('InputPrompt', () => {
 
       // 1. Verify initial placeholder
       await waitFor(() => {
-        expect(stdout.lastFrame()).toMatchSnapshot();
+        expect(stdout.lastFrame()).toContain('[Pasted Text: 10 lines]');
       });
+      expect(stdout.lastFrame()).toMatchSnapshot();
 
       // Simulate double-click to expand
       await simulateClick(5, 2);
@@ -3886,8 +3887,9 @@ describe('InputPrompt', () => {
 
       // 2. Verify expanded content is visible
       await waitFor(() => {
-        expect(stdout.lastFrame()).toMatchSnapshot();
+        expect(stdout.lastFrame()).toContain('line10');
       });
+      expect(stdout.lastFrame()).toMatchSnapshot();
 
       // Simulate double-click to collapse
       await simulateClick(5, 2);
@@ -3895,8 +3897,9 @@ describe('InputPrompt', () => {
 
       // 3. Verify placeholder is restored
       await waitFor(() => {
-        expect(stdout.lastFrame()).toMatchSnapshot();
+        expect(stdout.lastFrame()).toContain('[Pasted Text: 10 lines]');
       });
+      expect(stdout.lastFrame()).toMatchSnapshot();
 
       unmount();
     });

packages/cli/src/ui/components/__snapshots__/InputPrompt.test.tsx.snap

@@ -156,7 +156,16 @@ exports[`InputPrompt > mouse interaction > should toggle paste expansion on doub
 
 exports[`InputPrompt > mouse interaction > should toggle paste expansion on double-click 2`] = `
 "▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀
- > [Pasted Text: 10 lines]                                                                          
+ > line1                                                                                            
+   line2                                                                                            
+   line3                                                                                            
+   line4                                                                                            
+   line5                                                                                            
+   line6                                                                                            
+   line7                                                                                            
+   line8                                                                                            
+   line9                                                                                            
+   line10                                                                                           
 ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄
 "
 `;

packages/cli/src/ui/components/__snapshots__/ThemeDialog.test.tsx.snap

@@ -11,12 +11,12 @@ exports[`Initial Theme Selection > should default to a dark theme when terminal
 │ ●  4. Default Dark (Matches terminal)        │ 3     a, b = 0, 1                               │ │
 │    5. Dracula Dark                           │ 4     for _ in range(n):                        │ │
 │    6. GitHub Dark                            │ 5         a, b = b, a + b                       │ │
-│    7. Holiday Dark                           │ 6     return a                                  │ │
-│    8. Shades Of Purple Dark                  │                                                 │ │
-│    9. Solarized Dark                         │ 1 - print("Hello, " + name)                     │ │
-│   10. Tokyo Night Dark                       │ 1 + print(f"Hello, {name}!")                    │ │
-│   11. ANSI Light                             │                                                 │ │
-│   12. Ayu Light                              └─────────────────────────────────────────────────┘ │
+│    7. GitHub Dark Colorblind Dark            │ 6     return a                                  │ │
+│    8. Holiday Dark                           │                                                 │ │
+│    9. Shades Of Purple Dark                  │ 1 - print("Hello, " + name)                     │ │
+│   10. Solarized Dark                         │ 1 + print(f"Hello, {name}!")                    │ │
+│   11. Tokyo Night Dark                       │                                                 │ │
+│   12. ANSI Light                             └─────────────────────────────────────────────────┘ │
 │ ▼                                                                                                │
 │                                                                                                  │
 │ (Use Enter to select, Tab to configure scope, Esc to close)                                      │
@@ -34,14 +34,14 @@ exports[`Initial Theme Selection > should default to a light theme when terminal
 │    2. Ayu Light                              │ 1 # function                                    │ │
 │ ●  3. Default Light                          │ 2 def fibonacci(n):                             │ │
 │    4. GitHub Light                           │ 3     a, b = 0, 1                               │ │
-│    5. Google Code Light                      │ 4     for _ in range(n):                        │ │
-│    6. Solarized Light                        │ 5         a, b = b, a + b                       │ │
-│    7. Xcode Light                            │ 6     return a                                  │ │
-│    8. ANSI Dark (Incompatible)               │                                                 │ │
-│    9. Atom One Dark (Incompatible)           │ 1 - print("Hello, " + name)                     │ │
-│   10. Ayu Dark (Incompatible)                │ 1 + print(f"Hello, {name}!")                    │ │
-│   11. Default Dark (Incompatible)            │                                                 │ │
-│   12. Dracula Dark (Incompatible)            └─────────────────────────────────────────────────┘ │
+│    5. GitHub Light Colorblind Light (Mat…    │ 4     for _ in range(n):                        │ │
+│    6. Google Code Light                      │ 5         a, b = b, a + b                       │ │
+│    7. Solarized Light                        │ 6     return a                                  │ │
+│    8. Xcode Light                            │                                                 │ │
+│    9. ANSI Dark (Incompatible)               │ 1 - print("Hello, " + name)                     │ │
+│   10. Atom One Dark (Incompatible)           │ 1 + print(f"Hello, {name}!")                    │ │
+│   11. Ayu Dark (Incompatible)                │                                                 │ │
+│   12. Default Dark (Incompatible)            └─────────────────────────────────────────────────┘ │
 │ ▼                                                                                                │
 │                                                                                                  │
 │ (Use Enter to select, Tab to configure scope, Esc to close)                                      │
@@ -61,12 +61,12 @@ exports[`Initial Theme Selection > should use the theme from settings even if te
 │    4. Default Dark (Matches terminal)        │ 3     a, b = 0, 1                               │ │
 │    5. Dracula Dark                           │ 4     for _ in range(n):                        │ │
 │    6. GitHub Dark                            │ 5         a, b = b, a + b                       │ │
-│    7. Holiday Dark                           │ 6     return a                                  │ │
-│    8. Shades Of Purple Dark                  │                                                 │ │
-│    9. Solarized Dark                         │ 1 - print("Hello, " + name)                     │ │
-│   10. Tokyo Night Dark                       │ 1 + print(f"Hello, {name}!")                    │ │
-│   11. ANSI Light                             │                                                 │ │
-│   12. Ayu Light                              └─────────────────────────────────────────────────┘ │
+│    7. GitHub Dark Colorblind Dark            │ 6     return a                                  │ │
+│    8. Holiday Dark                           │                                                 │ │
+│    9. Shades Of Purple Dark                  │ 1 - print("Hello, " + name)                     │ │
+│   10. Solarized Dark                         │ 1 + print(f"Hello, {name}!")                    │ │
+│   11. Tokyo Night Dark                       │                                                 │ │
+│   12. ANSI Light                             └─────────────────────────────────────────────────┘ │
 │ ▼                                                                                                │
 │                                                                                                  │
 │ (Use Enter to select, Tab to configure scope, Esc to close)                                      │
@@ -100,12 +100,12 @@ exports[`ThemeDialog Snapshots > should render correctly in theme selection mode
 │    4. Default Dark                           │ 3     a, b = 0, 1                               │ │
 │    5. Dracula Dark                           │ 4     for _ in range(n):                        │ │
 │    6. GitHub Dark                            │ 5         a, b = b, a + b                       │ │
-│    7. Holiday Dark                           │ 6     return a                                  │ │
-│    8. Shades Of Purple Dark                  │                                                 │ │
-│    9. Solarized Dark                         │ 1 - print("Hello, " + name)                     │ │
-│   10. Tokyo Night Dark                       │ 1 + print(f"Hello, {name}!")                    │ │
-│   11. ANSI Light                             │                                                 │ │
-│   12. Ayu Light                              └─────────────────────────────────────────────────┘ │
+│    7. GitHub Dark Colorblind Dark            │ 6     return a                                  │ │
+│    8. Holiday Dark                           │                                                 │ │
+│    9. Shades Of Purple Dark                  │ 1 - print("Hello, " + name)                     │ │
+│   10. Solarized Dark                         │ 1 + print(f"Hello, {name}!")                    │ │
+│   11. Tokyo Night Dark                       │                                                 │ │
+│   12. ANSI Light                             └─────────────────────────────────────────────────┘ │
 │ ▼                                                                                                │
 │                                                                                                  │
 │ (Use Enter to select, Tab to configure scope, Esc to close)                                      │
@@ -125,12 +125,12 @@ exports[`ThemeDialog Snapshots > should render correctly in theme selection mode
 │    4. Default Dark                           │ 3     a, b = 0, 1                               │ │
 │    5. Dracula Dark                           │ 4     for _ in range(n):                        │ │
 │    6. GitHub Dark                            │ 5         a, b = b, a + b                       │ │
-│    7. Holiday Dark                           │ 6     return a                                  │ │
-│    8. Shades Of Purple Dark                  │                                                 │ │
-│    9. Solarized Dark                         │ 1 - print("Hello, " + name)                     │ │
-│   10. Tokyo Night Dark                       │ 1 + print(f"Hello, {name}!")                    │ │
-│   11. ANSI Light                             │                                                 │ │
-│   12. Ayu Light                              └─────────────────────────────────────────────────┘ │
+│    7. GitHub Dark Colorblind Dark            │ 6     return a                                  │ │
+│    8. Holiday Dark                           │                                                 │ │
+│    9. Shades Of Purple Dark                  │ 1 - print("Hello, " + name)                     │ │
+│   10. Solarized Dark                         │ 1 + print(f"Hello, {name}!")                    │ │
+│   11. Tokyo Night Dark                       │                                                 │ │
+│   12. ANSI Light                             └─────────────────────────────────────────────────┘ │
 │ ▼                                                                                                │
 │                                              ╭─────────────────────────────────────────────────╮ │
 │                                              │ DEVELOPER TOOLS (Not visible to users)          │ │

packages/cli/src/ui/components/shared/BaseSelectionList.test.tsx

@@ -4,7 +4,8 @@
  * SPDX-License-Identifier: Apache-2.0
  */
 
-import { describe, it, expect, vi, beforeEach } from 'vitest';
+import { describe, it, expect, vi, beforeEach, type Mock } from 'vitest';
+import { act } from 'react';
 import { renderWithProviders } from '../../../test-utils/render.js';
 import {
   BaseSelectionList,
@@ -14,8 +15,10 @@ import {
 import { useSelectionList } from '../../hooks/useSelectionList.js';
 import { Text } from 'ink';
 import type { theme } from '../../semantic-colors.js';
+import { useMouseClick } from '../../hooks/useMouseClick.js';
 
 vi.mock('../../hooks/useSelectionList.js');
+vi.mock('../../hooks/useMouseClick.js');
 
 const mockTheme = {
   text: { primary: 'COLOR_PRIMARY', secondary: 'COLOR_SECONDARY' },
@@ -35,6 +38,7 @@ describe('BaseSelectionList', () => {
   const mockOnSelect = vi.fn();
   const mockOnHighlight = vi.fn();
   const mockRenderItem = vi.fn();
+  const mockSetActiveIndex = vi.fn();
 
   const items = [
     { value: 'A', label: 'Item A', key: 'A' },
@@ -54,7 +58,7 @@ describe('BaseSelectionList', () => {
   ) => {
     vi.mocked(useSelectionList).mockReturnValue({
       activeIndex,
-      setActiveIndex: vi.fn(),
+      setActiveIndex: mockSetActiveIndex,
     });
 
     mockRenderItem.mockImplementation(
@@ -484,6 +488,79 @@ describe('BaseSelectionList', () => {
     });
   });
 
+  describe('Mouse Interaction', () => {
+    it('should register mouse click handler for each item', async () => {
+      const { unmount } = await renderComponent();
+
+      // items are A, B (disabled), C
+      expect(useMouseClick).toHaveBeenCalledTimes(3);
+      unmount();
+    });
+
+    it('should update activeIndex on first click and call onSelect on second click', async () => {
+      const { unmount, waitUntilReady } = await renderComponent();
+      await waitUntilReady();
+
+      // items[0] is 'A' (enabled)
+      // items[1] is 'B' (disabled)
+      // items[2] is 'C' (enabled)
+
+      // Get the mouse click handler for the third item (index 2)
+      const mouseClickHandler = (useMouseClick as Mock).mock.calls[2][1];
+
+      // First click on item C
+      act(() => {
+        mouseClickHandler();
+      });
+
+      expect(mockSetActiveIndex).toHaveBeenCalledWith(2);
+      expect(mockOnSelect).not.toHaveBeenCalled();
+
+      // Now simulate being on item C (isSelected = true)
+      // Rerender or update mocks for the next check
+      await renderComponent({}, 2);
+
+      // Get the updated mouse click handler for item C
+      // useMouseClick is called 3 more times on rerender
+      const updatedMouseClickHandler = (useMouseClick as Mock).mock.calls[5][1];
+
+      // Second click on item C
+      act(() => {
+        updatedMouseClickHandler();
+      });
+
+      expect(mockOnSelect).toHaveBeenCalledWith('C');
+      unmount();
+    });
+
+    it('should not call onSelect when a disabled item is clicked', async () => {
+      const { unmount, waitUntilReady } = await renderComponent();
+      await waitUntilReady();
+
+      // items[1] is 'B' (disabled)
+      const mouseClickHandler = (useMouseClick as Mock).mock.calls[1][1];
+
+      act(() => {
+        mouseClickHandler();
+      });
+
+      expect(mockSetActiveIndex).not.toHaveBeenCalled();
+      expect(mockOnSelect).not.toHaveBeenCalled();
+      unmount();
+    });
+
+    it('should pass isActive: isFocused to useMouseClick', async () => {
+      const { unmount } = await renderComponent({ isFocused: false });
+
+      expect(useMouseClick).toHaveBeenCalledWith(
+        expect.any(Object),
+        expect.any(Function),
+        { isActive: false },
+      );
+      unmount();
+    });
+  });
+
   describe('Scroll Arrows (showScrollArrows)', () => {
     const longList = Array.from({ length: 10 }, (_, i) => ({
       value: `Item ${i + 1}`,

packages/cli/src/ui/components/shared/BaseSelectionList.tsx

@@ -5,13 +5,14 @@
  */
 
 import type React from 'react';
-import { useState } from 'react';
-import { Text, Box } from 'ink';
+import { useState, useRef } from 'react';
+import { Text, Box, type DOMElement } from 'ink';
 import { theme } from '../../semantic-colors.js';
 import {
   useSelectionList,
   type SelectionListItem,
 } from '../../hooks/useSelectionList.js';
+import { useMouseClick } from '../../hooks/useMouseClick.js';
 
 export interface RenderItemContext {
   isSelected: boolean;
@@ -38,6 +39,119 @@ export interface BaseSelectionListProps<
   renderItem: (item: TItem, context: RenderItemContext) => React.ReactNode;
 }
 
+interface SelectionListItemRowProps<
+  T,
+  TItem extends SelectionListItem<T> = SelectionListItem<T>,
+> {
+  item: TItem;
+  itemIndex: number;
+  isSelected: boolean;
+  isFocused: boolean;
+  showNumbers: boolean;
+  selectedIndicator: string;
+  numberColumnWidth: number;
+  onSelect: (value: T) => void;
+  setActiveIndex: (index: number) => void;
+  renderItem: (item: TItem, context: RenderItemContext) => React.ReactNode;
+}
+
+function SelectionListItemRow<
+  T,
+  TItem extends SelectionListItem<T> = SelectionListItem<T>,
+>({
+  item,
+  itemIndex,
+  isSelected,
+  isFocused,
+  showNumbers,
+  selectedIndicator,
+  numberColumnWidth,
+  onSelect,
+  setActiveIndex,
+  renderItem,
+}: SelectionListItemRowProps<T, TItem>) {
+  const containerRef = useRef<DOMElement>(null);
+
+  useMouseClick(
+    containerRef,
+    () => {
+      if (!item.disabled) {
+        if (isSelected) {
+          // Second click on the same item triggers submission
+          onSelect(item.value);
+        } else {
+          // First click highlights the item
+          setActiveIndex(itemIndex);
+        }
+      }
+    },
+    { isActive: isFocused },
+  );
+
+  let titleColor = theme.text.primary;
+  let numberColor = theme.text.primary;
+
+  if (isSelected) {
+    titleColor = theme.ui.focus;
+    numberColor = theme.ui.focus;
+  } else if (item.disabled) {
+    titleColor = theme.text.secondary;
+    numberColor = theme.text.secondary;
+  }
+
+  if (!isFocused && !item.disabled) {
+    numberColor = theme.text.secondary;
+  }
+
+  if (!showNumbers) {
+    numberColor = theme.text.secondary;
+  }
+
+  const itemNumberText = `${String(itemIndex + 1).padStart(
+    numberColumnWidth,
+  )}.`;
+
+  return (
+    <Box
+      ref={containerRef}
+      key={item.key}
+      alignItems="flex-start"
+      backgroundColor={isSelected ? theme.background.focus : undefined}
+    >
+      {/* Radio button indicator */}
+      <Box minWidth={2} flexShrink={0}>
+        <Text
+          color={isSelected ? theme.ui.focus : theme.text.primary}
+          aria-hidden
+        >
+          {isSelected ? selectedIndicator : ' '}
+        </Text>
+      </Box>
+
+      {/* Item number */}
+      {showNumbers && !item.hideNumber && (
+        <Box
+          marginRight={1}
+          flexShrink={0}
+          minWidth={itemNumberText.length}
+          aria-state={{ checked: isSelected }}
+        >
+          <Text color={numberColor}>{itemNumberText}</Text>
+        </Box>
+      )}
+
+      {/* Custom content via render prop */}
+      <Box flexGrow={1}>
+        {renderItem(item, {
+          isSelected,
+          titleColor,
+          numberColor,
+        })}
+      </Box>
+    </Box>
+  );
+}
+
 /**
  * Base component for selection lists that provides common UI structure
  * and keyboard navigation logic via the useSelectionList hook.
@@ -70,7 +184,7 @@ export function BaseSelectionList<
   selectedIndicator = '●',
   renderItem,
 }: BaseSelectionListProps<T, TItem>): React.JSX.Element {
-  const { activeIndex } = useSelectionList({
+  const { activeIndex, setActiveIndex } = useSelectionList({
     items,
     initialIndex,
     onSelect,
@@ -107,10 +221,12 @@ export function BaseSelectionList<
   );
   const numberColumnWidth = String(items.length).length;
 
+  const showArrows = showScrollArrows && items.length > maxItemsToShow;
+
   return (
     <Box flexDirection="column">
       {/* Use conditional coloring instead of conditional rendering */}
-      {showScrollArrows && items.length > maxItemsToShow && (
+      {showArrows && (
         <Text
           color={
             effectiveScrollOffset > 0
@@ -126,71 +242,24 @@ export function BaseSelectionList<
         const itemIndex = effectiveScrollOffset + index;
         const isSelected = activeIndex === itemIndex;
 
-        // Determine colors based on selection and disabled state
-        let titleColor = theme.text.primary;
-        let numberColor = theme.text.primary;
-
-        if (isSelected) {
-          titleColor = theme.ui.focus;
-          numberColor = theme.ui.focus;
-        } else if (item.disabled) {
-          titleColor = theme.text.secondary;
-          numberColor = theme.text.secondary;
-        }
-
-        if (!isFocused && !item.disabled) {
-          numberColor = theme.text.secondary;
-        }
-
-        if (!showNumbers) {
-          numberColor = theme.text.secondary;
-        }
-
-        const itemNumberText = `${String(itemIndex + 1).padStart(
-          numberColumnWidth,
-        )}.`;
-
         return (
-          <Box
+          <SelectionListItemRow
             key={item.key}
-            alignItems="flex-start"
-            backgroundColor={isSelected ? theme.background.focus : undefined}
-          >
-            {/* Radio button indicator */}
-            <Box minWidth={2} flexShrink={0}>
-              <Text
-                color={isSelected ? theme.ui.focus : theme.text.primary}
-                aria-hidden
-              >
-                {isSelected ? selectedIndicator : ' '}
-              </Text>
-            </Box>
-
-            {/* Item number */}
-            {showNumbers && !item.hideNumber && (
-              <Box
-                marginRight={1}
-                flexShrink={0}
-                minWidth={itemNumberText.length}
-                aria-state={{ checked: isSelected }}
-              >
-                <Text color={numberColor}>{itemNumberText}</Text>
-              </Box>
-            )}
-
-            {/* Custom content via render prop */}
-            <Box flexGrow={1}>
-              {renderItem(item, {
-                isSelected,
-                titleColor,
-                numberColor,
-              })}
-            </Box>
-          </Box>
+            item={item}
+            itemIndex={itemIndex}
+            isSelected={isSelected}
+            isFocused={isFocused}
+            showNumbers={showNumbers}
+            selectedIndicator={selectedIndicator}
+            numberColumnWidth={numberColumnWidth}
+            onSelect={onSelect}
+            setActiveIndex={setActiveIndex}
+            renderItem={renderItem}
+          />
         );
       })}
 
-      {showScrollArrows && items.length > maxItemsToShow && (
+      {showArrows && (
         <Text
           color={
             effectiveScrollOffset + maxItemsToShow < items.length

packages/cli/src/ui/components/shared/TextInput.test.tsx

@@ -11,12 +11,17 @@ import { act } from 'react';
 import { TextInput } from './TextInput.js';
 import { useKeypress } from '../../hooks/useKeypress.js';
 import { useTextBuffer, type TextBuffer } from './text-buffer.js';
+import { useMouseClick } from '../../hooks/useMouseClick.js';
 
 // Mocks
 vi.mock('../../hooks/useKeypress.js', () => ({
   useKeypress: vi.fn(),
 }));
 
+vi.mock('../../hooks/useMouseClick.js', () => ({
+  useMouseClick: vi.fn(),
+}));
+
 vi.mock('./text-buffer.js', async (importOriginal) => {
   const actual = await importOriginal<typeof import('./text-buffer.js')>();
   const mockTextBuffer = {
@@ -69,6 +74,7 @@ vi.mock('./text-buffer.js', async (importOriginal) => {
 
 const mockedUseKeypress = useKeypress as Mock;
 const mockedUseTextBuffer = useTextBuffer as Mock;
+const mockedUseMouseClick = useMouseClick as Mock;
 
 describe('TextInput', () => {
   const onCancel = vi.fn();
@@ -84,6 +90,7 @@ describe('TextInput', () => {
       cursor: [0, 0],
       visualCursor: [0, 0],
       viewportVisualLines: [''],
+      visualScrollRow: 0,
       pastedContent: {} as Record<string, string>,
       handleInput: vi.fn((key) => {
         if (key.sequence) {
@@ -408,4 +415,36 @@ describe('TextInput', () => {
     expect(lastFrame()).toContain('line2');
     unmount();
   });
+
+  it('registers mouse click handler for free-form text input', async () => {
+    const { unmount } = await render(
+      <TextInput buffer={mockBuffer} onSubmit={onSubmit} onCancel={onCancel} />,
+    );
+
+    expect(mockedUseMouseClick).toHaveBeenCalledWith(
+      expect.any(Object),
+      expect.any(Function),
+      expect.objectContaining({ isActive: true, name: 'left-press' }),
+    );
+    unmount();
+  });
+
+  it('registers mouse click handler for placeholder view', async () => {
+    mockBuffer.text = '';
+    const { unmount } = await render(
+      <TextInput
+        buffer={mockBuffer}
+        placeholder="test"
+        onSubmit={onSubmit}
+        onCancel={onCancel}
+      />,
+    );
+
+    expect(mockedUseMouseClick).toHaveBeenCalledWith(
+      expect.any(Object),
+      expect.any(Function),
+      expect.objectContaining({ isActive: true, name: 'left-press' }),
+    );
+    unmount();
+  });
 });

packages/cli/src/ui/components/shared/TextInput.tsx

@@ -5,8 +5,8 @@
  */
 
 import type React from 'react';
-import { useCallback } from 'react';
-import { Text, Box } from 'ink';
+import { useCallback, useRef } from 'react';
+import { Text, Box, type DOMElement } from 'ink';
 import { useKeypress, type Key } from '../../hooks/useKeypress.js';
 import chalk from 'chalk';
 import { theme } from '../../semantic-colors.js';
@@ -14,6 +14,7 @@ import { expandPastePlaceholders, type TextBuffer } from './text-buffer.js';
 import { cpSlice, cpIndexToOffset } from '../../utils/textUtils.js';
 import { Command } from '../../key/keyMatchers.js';
 import { useKeyMatchers } from '../../hooks/useKeyMatchers.js';
+import { useMouseClick } from '../../hooks/useMouseClick.js';
 
 export interface TextInputProps {
   buffer: TextBuffer;
@@ -31,6 +32,8 @@ export function TextInput({
   focus = true,
 }: TextInputProps): React.JSX.Element {
   const keyMatchers = useKeyMatchers();
+  const containerRef = useRef<DOMElement>(null);
+
   const {
     text,
     handleInput,
@@ -40,6 +43,17 @@ export function TextInput({
   } = buffer;
   const [cursorVisualRowAbsolute, cursorVisualColAbsolute] = visualCursor;
 
+  useMouseClick(
+    containerRef,
+    (_event, relativeX, relativeY) => {
+      if (focus) {
+        const visRowAbsolute = visualScrollRow + relativeY;
+        buffer.moveToVisualPosition(visRowAbsolute, relativeX);
+      }
+    },
+    { isActive: focus, name: 'left-press' },
+  );
+
   const handleKeyPress = useCallback(
     (key: Key) => {
       if (key.name === 'escape' && onCancel) {
@@ -64,7 +78,7 @@ export function TextInput({
 
   if (showPlaceholder) {
     return (
-      <Box>
+      <Box ref={containerRef}>
         {focus ? (
           <Text terminalCursorFocus={focus} terminalCursorPosition={0}>
             {chalk.inverse(placeholder[0] || ' ')}
@@ -78,7 +92,7 @@ export function TextInput({
   }
 
   return (
-    <Box flexDirection="column">
+    <Box ref={containerRef} flexDirection="column">
       {viewportVisualLines.map((lineText, idx) => {
         const currentVisualRow = visualScrollRow + idx;
         const isCursorLine =

packages/cli/src/ui/hooks/slashCommandProcessor.test.tsx

@@ -858,11 +858,81 @@ describe('useSlashCommandProcessor', () => {
   });
 
   describe('Lifecycle', () => {
+    it('removes the IDE status listener on unmount after async initialization', async () => {
+      let resolveIdeClient:
+        | ((client: {
+            addStatusChangeListener: (listener: () => void) => void;
+            removeStatusChangeListener: (listener: () => void) => void;
+          }) => void)
+        | undefined;
+      const addStatusChangeListener = vi.fn();
+      const removeStatusChangeListener = vi.fn();
+
+      mockIdeClientGetInstance.mockImplementation(
+        () =>
+          new Promise((resolve) => {
+            resolveIdeClient = resolve;
+          }),
+      );
+
+      const result = await setupProcessorHook();
+
+      await act(async () => {
+        resolveIdeClient?.({
+          addStatusChangeListener,
+          removeStatusChangeListener,
+        });
+      });
+
+      result.unmount();
+      unmountHook = undefined;
+
+      expect(addStatusChangeListener).toHaveBeenCalledTimes(1);
+      expect(removeStatusChangeListener).toHaveBeenCalledTimes(1);
+      expect(removeStatusChangeListener).toHaveBeenCalledWith(
+        addStatusChangeListener.mock.calls[0]?.[0],
+      );
+    });
+
+    it('does not register an IDE status listener if unmounted before async initialization resolves', async () => {
+      let resolveIdeClient:
+        | ((client: {
+            addStatusChangeListener: (listener: () => void) => void;
+            removeStatusChangeListener: (listener: () => void) => void;
+          }) => void)
+        | undefined;
+      const addStatusChangeListener = vi.fn();
+      const removeStatusChangeListener = vi.fn();
+
+      mockIdeClientGetInstance.mockImplementation(
+        () =>
+          new Promise((resolve) => {
+            resolveIdeClient = resolve;
+          }),
+      );
+
+      const result = await setupProcessorHook();
+
+      result.unmount();
+      unmountHook = undefined;
+
+      await act(async () => {
+        resolveIdeClient?.({
+          addStatusChangeListener,
+          removeStatusChangeListener,
+        });
+      });
+
+      expect(addStatusChangeListener).not.toHaveBeenCalled();
+      expect(removeStatusChangeListener).not.toHaveBeenCalled();
+    });
+
     it('should abort command loading when the hook unmounts', async () => {
       const abortSpy = vi.spyOn(AbortController.prototype, 'abort');
       const { unmount } = await setupProcessorHook();
 
       unmount();
+      unmountHook = undefined;
 
       expect(abortSpy).toHaveBeenCalledTimes(1);
     });

packages/cli/src/ui/hooks/slashCommandProcessor.ts

@@ -281,10 +281,16 @@ export const useSlashCommandProcessor = (
     const listener = () => {
       reloadCommands();
     };
+    let isActive = true;
+    let activeIdeClient: IdeClient | undefined;
 
     // eslint-disable-next-line @typescript-eslint/no-floating-promises
     (async () => {
       const ideClient = await IdeClient.getInstance();
+      if (!isActive) {
+        return;
+      }
+      activeIdeClient = ideClient;
       ideClient.addStatusChangeListener(listener);
     })();
 
@@ -307,11 +313,8 @@ export const useSlashCommandProcessor = (
     coreEvents.on('extensionsStopping', extensionEventListener);
 
     return () => {
-      // eslint-disable-next-line @typescript-eslint/no-floating-promises
-      (async () => {
-        const ideClient = await IdeClient.getInstance();
-        ideClient.removeStatusChangeListener(listener);
-      })();
+      isActive = false;
+      activeIdeClient?.removeStatusChangeListener(listener);
       removeMCPStatusChangeListener(listener);
       coreEvents.off('extensionsStarting', extensionEventListener);
       coreEvents.off('extensionsStopping', extensionEventListener);

packages/cli/src/ui/hooks/useExecutionLifecycle.test.tsx

@@ -76,7 +76,23 @@ vi.mock('@google/gemini-cli-core', async (importOriginal) => {
     isBinary: mockIsBinary,
   };
 });
-vi.mock('node:fs');
+vi.mock('node:fs', async (importOriginal) => {
+  const actual = await importOriginal<typeof import('node:fs')>();
+  const mocked = {
+    ...actual,
+    existsSync: vi.fn(),
+    readFileSync: vi.fn(),
+    createWriteStream: vi.fn(),
+    promises: {
+      ...actual.promises,
+      unlink: vi.fn().mockResolvedValue(undefined),
+    },
+  };
+  return {
+    ...mocked,
+    default: mocked,
+  };
+});
 vi.mock('node:os', async (importOriginal) => {
   const actual = await importOriginal<typeof import('node:os')>();
   const mocked = {
@@ -144,6 +160,10 @@ describe('useExecutionLifecycle', () => {
           enableEnvironmentVariableRedaction: false,
         },
       }),
+      getTruncateToolOutputThreshold: () => 40000,
+      storage: {
+        getProjectTempDir: () => '/tmp/project',
+      },
     } as unknown as Config;
     mockGeminiClient = { addHistory: vi.fn() } as unknown as GeminiClient;
 
@@ -155,6 +175,16 @@ describe('useExecutionLifecycle', () => {
     mockIsBinary.mockReturnValue(false);
     vi.mocked(fs.existsSync).mockReturnValue(false);
 
+    vi.mocked(fs.createWriteStream).mockReturnValue({
+      write: vi.fn(),
+      end: vi.fn().mockImplementation((cb: () => void) => {
+        if (cb) cb();
+      }),
+      destroy: vi.fn(),
+      bytesWritten: 0,
+      closed: false,
+    } as unknown as fs.WriteStream);
+
     mockShellExecutionService.mockImplementation((_cmd, _cwd, callback) => {
       mockShellOutputCallback = callback;
       return Promise.resolve({
@@ -646,7 +676,7 @@ describe('useExecutionLifecycle', () => {
     });
     const tmpFile = path.join(os.tmpdir(), 'shell_pwd_abcdef.tmp');
     // Verify that the temporary file was cleaned up
-    expect(vi.mocked(fs.unlinkSync)).toHaveBeenCalledWith(tmpFile);
+    expect(vi.mocked(fs.promises.unlink)).toHaveBeenCalledWith(tmpFile);
     expect(setShellInputFocusedMock).toHaveBeenCalledWith(false);
   });
 
@@ -674,7 +704,7 @@ describe('useExecutionLifecycle', () => {
       expect(finalHistoryItem.tools[0].resultDisplay).toContain(
         "WARNING: shell mode is stateless; the directory change to '/test/dir/new' will not persist.",
       );
-      expect(vi.mocked(fs.unlinkSync)).toHaveBeenCalledWith(tmpFile);
+      expect(vi.mocked(fs.promises.unlink)).toHaveBeenCalledWith(tmpFile);
     });
 
     it('should NOT show a warning if the directory does not change', async () => {

packages/cli/src/ui/hooks/useExecutionLifecycle.ts

@@ -20,6 +20,8 @@ import {
   ShellExecutionService,
   ExecutionLifecycleService,
   CoreToolCallStatus,
+  moveToolOutputToFile,
+  debugLogger,
 } from '@google/gemini-cli-core';
 import { type PartListUnion } from '@google/genai';
 import type { UseHistoryManagerReturn } from './useHistoryManager.js';
@@ -39,16 +41,15 @@ export { type BackgroundTask };
 
 export const OUTPUT_UPDATE_INTERVAL_MS = 1000;
 const RESTORE_VISIBILITY_DELAY_MS = 300;
-const MAX_OUTPUT_LENGTH = 10000;
-
 function addShellCommandToGeminiHistory(
   geminiClient: GeminiClient,
   rawQuery: string,
   resultText: string,
+  maxOutputLength: number,
 ) {
   const modelContent =
-    resultText.length > MAX_OUTPUT_LENGTH
-      ? resultText.substring(0, MAX_OUTPUT_LENGTH) + '\n... (truncated)'
+    maxOutputLength > 0 && resultText.length > maxOutputLength
+      ? resultText.substring(0, maxOutputLength) + '\n... (truncated)'
       : resultText;
 
   // Escape backticks to prevent prompt injection breakouts
@@ -424,6 +425,9 @@ export const useExecutionLifecycle = (
                 let shouldUpdate = false;
 
                 switch (event.type) {
+                  case 'raw_data':
+                  case 'file_data':
+                    break;
                   case 'data':
                     if (isBinaryStream) break;
                     if (typeof event.chunk === 'string') {
@@ -533,6 +537,24 @@ export const useExecutionLifecycle = (
           } else {
             mainContent =
               result.output.trim() || '(Command produced no output)';
+            if (result.fullOutputFilePath) {
+              const { outputFile: savedPath } = await moveToolOutputToFile(
+                result.fullOutputFilePath,
+                SHELL_COMMAND_NAME,
+                callId,
+                config.storage.getProjectTempDir(),
+                config.getSessionId(),
+              );
+              const warning = `[Full command output saved to: ${savedPath}]`;
+              mainContent = mainContent.includes(
+                '[GEMINI_CLI_WARNING: Output truncated.',
+              )
+                ? mainContent.replace(
+                    /\[GEMINI_CLI_WARNING: Output truncated\..*?\]/,
+                    warning,
+                  )
+                : `${mainContent}\n\n${warning}`;
+            }
           }
 
           let finalOutput: string | AnsiOutput =
@@ -617,7 +639,12 @@ export const useExecutionLifecycle = (
             );
           }
 
-          addShellCommandToGeminiHistory(geminiClient, rawQuery, mainContent);
+          addShellCommandToGeminiHistory(
+            geminiClient,
+            rawQuery,
+            mainContent,
+            config.getTruncateToolOutputThreshold(),
+          );
         } catch (err) {
           setPendingHistoryItem(null);
           const errorMessage = err instanceof Error ? err.message : String(err);
@@ -630,8 +657,13 @@ export const useExecutionLifecycle = (
           );
         } finally {
           abortSignal.removeEventListener('abort', abortHandler);
-          if (pwdFilePath && fs.existsSync(pwdFilePath)) {
-            fs.unlinkSync(pwdFilePath);
+          if (pwdFilePath) {
+            fs.promises.unlink(pwdFilePath).catch((err) => {
+              debugLogger.warn(
+                `Failed to cleanup pwd file: ${pwdFilePath}`,
+                err,
+              );
+            });
           }
 
           dispatch({ type: 'SET_ACTIVE_PTY', pid: null });

packages/cli/src/ui/themes/builtin/dark/github-dark-colorblind.ts

@@ -0,0 +1,147 @@
+/**
+ * @license
+ * Copyright 2025 Google LLC
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+import { type ColorsTheme, Theme } from '../../theme.js';
+import { interpolateColor } from '../../color-utils.js';
+
+const githubDarkColorblindColors: ColorsTheme = {
+  type: 'dark',
+  Background: '#0d1117',
+  Foreground: '#e6edf3',
+  LightBlue: '#a5d6ff',
+  AccentBlue: '#79c0ff',
+  AccentPurple: '#d2a8ff',
+  AccentCyan: '#a5d6ff',
+  AccentGreen: '#a5d6ff',
+  AccentYellow: '#d29922',
+  AccentRed: '#f0883e',
+  DiffAdded: '#0d161f',
+  DiffRemoved: '#1d150e',
+  Comment: '#7d8590',
+  Gray: '#7d8590',
+  DarkGray: interpolateColor('#7d8590', '#0d1117', 0.5),
+  GradientColors: ['#58a6ff', '#f0883e'],
+};
+
+export const GitHubDarkColorblind: Theme = new Theme(
+  'GitHub Dark Colorblind',
+  'dark',
+  {
+    hljs: {
+      display: 'block',
+      overflowX: 'auto',
+      padding: '0.5em',
+      color: githubDarkColorblindColors.Foreground,
+      background: githubDarkColorblindColors.Background,
+    },
+    'hljs-comment': {
+      color: githubDarkColorblindColors.Comment,
+      fontStyle: 'italic',
+    },
+    'hljs-quote': {
+      color: githubDarkColorblindColors.Comment,
+      fontStyle: 'italic',
+    },
+    'hljs-keyword': {
+      color: githubDarkColorblindColors.AccentRed,
+      fontWeight: 'bold',
+    },
+    'hljs-selector-tag': {
+      color: githubDarkColorblindColors.AccentRed,
+      fontWeight: 'bold',
+    },
+    'hljs-subst': {
+      color: githubDarkColorblindColors.Foreground,
+    },
+    'hljs-number': {
+      color: githubDarkColorblindColors.LightBlue,
+    },
+    'hljs-literal': {
+      color: githubDarkColorblindColors.LightBlue,
+    },
+    'hljs-variable': {
+      color: githubDarkColorblindColors.Foreground,
+    },
+    'hljs-template-variable': {
+      color: githubDarkColorblindColors.Foreground,
+    },
+    'hljs-tag .hljs-attr': {
+      color: githubDarkColorblindColors.AccentYellow,
+    },
+    'hljs-string': {
+      color: githubDarkColorblindColors.AccentCyan,
+    },
+    'hljs-doctag': {
+      color: githubDarkColorblindColors.AccentCyan,
+    },
+    'hljs-title': {
+      color: githubDarkColorblindColors.AccentPurple,
+      fontWeight: 'bold',
+    },
+    'hljs-section': {
+      color: githubDarkColorblindColors.AccentPurple,
+      fontWeight: 'bold',
+    },
+    'hljs-selector-id': {
+      color: githubDarkColorblindColors.AccentPurple,
+      fontWeight: 'bold',
+    },
+    'hljs-type': {
+      color: githubDarkColorblindColors.AccentGreen,
+      fontWeight: 'bold',
+    },
+    'hljs-class .hljs-title': {
+      color: githubDarkColorblindColors.AccentGreen,
+      fontWeight: 'bold',
+    },
+    'hljs-tag': {
+      color: githubDarkColorblindColors.AccentGreen,
+    },
+    'hljs-name': {
+      color: githubDarkColorblindColors.AccentGreen,
+    },
+    'hljs-attribute': {
+      color: githubDarkColorblindColors.LightBlue,
+    },
+    'hljs-regexp': {
+      color: githubDarkColorblindColors.AccentCyan,
+    },
+    'hljs-link': {
+      color: githubDarkColorblindColors.AccentCyan,
+    },
+    'hljs-symbol': {
+      color: githubDarkColorblindColors.AccentPurple,
+    },
+    'hljs-bullet': {
+      color: githubDarkColorblindColors.AccentPurple,
+    },
+    'hljs-built_in': {
+      color: githubDarkColorblindColors.LightBlue,
+    },
+    'hljs-builtin-name': {
+      color: githubDarkColorblindColors.LightBlue,
+    },
+    'hljs-meta': {
+      color: githubDarkColorblindColors.LightBlue,
+      fontWeight: 'bold',
+    },
+    'hljs-deletion': {
+      background: '#682d0f',
+      color: githubDarkColorblindColors.AccentRed,
+    },
+    'hljs-addition': {
+      background: '#0c2d6b',
+      color: githubDarkColorblindColors.AccentGreen,
+    },
+    'hljs-emphasis': {
+      fontStyle: 'italic',
+    },
+    'hljs-strong': {
+      fontWeight: 'bold',
+    },
+  },
+  githubDarkColorblindColors,
+);

packages/cli/src/ui/themes/builtin/light/github-light-colorblind.ts

@@ -0,0 +1,147 @@
+/**
+ * @license
+ * Copyright 2025 Google LLC
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+import { type ColorsTheme, Theme } from '../../theme.js';
+import { interpolateColor } from '../../color-utils.js';
+
+const githubLightColorblindColors: ColorsTheme = {
+  type: 'light',
+  Background: '#ffffff',
+  Foreground: '#1f2328',
+  LightBlue: '#0a3069',
+  AccentBlue: '#0550ae',
+  AccentPurple: '#8250df',
+  AccentCyan: '#0a3069',
+  AccentGreen: '#0969da',
+  AccentYellow: '#9a6700',
+  AccentRed: '#bc4c00',
+  DiffAdded: '#ddf4ff',
+  DiffRemoved: '#fff1e5',
+  Comment: '#656d76',
+  Gray: '#656d76',
+  DarkGray: interpolateColor('#656d76', '#ffffff', 0.5),
+  GradientColors: ['#0969da', '#bc4c00'],
+};
+
+export const GitHubLightColorblind: Theme = new Theme(
+  'GitHub Light Colorblind',
+  'light',
+  {
+    hljs: {
+      display: 'block',
+      overflowX: 'auto',
+      padding: '0.5em',
+      color: githubLightColorblindColors.Foreground,
+      background: githubLightColorblindColors.Background,
+    },
+    'hljs-comment': {
+      color: githubLightColorblindColors.Comment,
+      fontStyle: 'italic',
+    },
+    'hljs-quote': {
+      color: githubLightColorblindColors.Comment,
+      fontStyle: 'italic',
+    },
+    'hljs-keyword': {
+      color: githubLightColorblindColors.AccentRed,
+      fontWeight: 'bold',
+    },
+    'hljs-selector-tag': {
+      color: githubLightColorblindColors.AccentRed,
+      fontWeight: 'bold',
+    },
+    'hljs-subst': {
+      color: githubLightColorblindColors.Foreground,
+    },
+    'hljs-number': {
+      color: githubLightColorblindColors.LightBlue,
+    },
+    'hljs-literal': {
+      color: githubLightColorblindColors.LightBlue,
+    },
+    'hljs-variable': {
+      color: githubLightColorblindColors.Foreground,
+    },
+    'hljs-template-variable': {
+      color: githubLightColorblindColors.Foreground,
+    },
+    'hljs-tag .hljs-attr': {
+      color: githubLightColorblindColors.AccentYellow,
+    },
+    'hljs-string': {
+      color: githubLightColorblindColors.AccentCyan,
+    },
+    'hljs-doctag': {
+      color: githubLightColorblindColors.AccentCyan,
+    },
+    'hljs-title': {
+      color: githubLightColorblindColors.AccentPurple,
+      fontWeight: 'bold',
+    },
+    'hljs-section': {
+      color: githubLightColorblindColors.AccentPurple,
+      fontWeight: 'bold',
+    },
+    'hljs-selector-id': {
+      color: githubLightColorblindColors.AccentPurple,
+      fontWeight: 'bold',
+    },
+    'hljs-type': {
+      color: githubLightColorblindColors.AccentGreen,
+      fontWeight: 'bold',
+    },
+    'hljs-class .hljs-title': {
+      color: githubLightColorblindColors.AccentGreen,
+      fontWeight: 'bold',
+    },
+    'hljs-tag': {
+      color: githubLightColorblindColors.AccentGreen,
+    },
+    'hljs-name': {
+      color: githubLightColorblindColors.AccentGreen,
+    },
+    'hljs-attribute': {
+      color: githubLightColorblindColors.LightBlue,
+    },
+    'hljs-regexp': {
+      color: githubLightColorblindColors.AccentCyan,
+    },
+    'hljs-link': {
+      color: githubLightColorblindColors.AccentCyan,
+    },
+    'hljs-symbol': {
+      color: githubLightColorblindColors.AccentPurple,
+    },
+    'hljs-bullet': {
+      color: githubLightColorblindColors.AccentPurple,
+    },
+    'hljs-built_in': {
+      color: githubLightColorblindColors.LightBlue,
+    },
+    'hljs-builtin-name': {
+      color: githubLightColorblindColors.LightBlue,
+    },
+    'hljs-meta': {
+      color: githubLightColorblindColors.LightBlue,
+      fontWeight: 'bold',
+    },
+    'hljs-deletion': {
+      background: '#fff1e5',
+      color: githubLightColorblindColors.AccentRed,
+    },
+    'hljs-addition': {
+      background: '#ddf4ff',
+      color: githubLightColorblindColors.AccentGreen,
+    },
+    'hljs-emphasis': {
+      fontStyle: 'italic',
+    },
+    'hljs-strong': {
+      fontWeight: 'bold',
+    },
+  },
+  githubLightColorblindColors,
+);

packages/cli/src/ui/themes/theme-manager.ts

@@ -10,6 +10,8 @@ import { AtomOneDark } from './builtin/dark/atom-one-dark.js';
 import { Dracula } from './builtin/dark/dracula-dark.js';
 import { GitHubDark } from './builtin/dark/github-dark.js';
 import { GitHubLight } from './builtin/light/github-light.js';
+import { GitHubDarkColorblind } from './builtin/dark/github-dark-colorblind.js';
+import { GitHubLightColorblind } from './builtin/light/github-light-colorblind.js';
 import { GoogleCode } from './builtin/light/googlecode-light.js';
 import { Holiday } from './builtin/dark/holiday-dark.js';
 import { DefaultLight } from './builtin/light/default-light.js';
@@ -79,6 +81,8 @@ class ThemeManager {
       DefaultDark,
       GitHubDark,
       GitHubLight,
+      GitHubDarkColorblind,
+      GitHubLightColorblind,
       GoogleCode,
       Holiday,
       ShadesOfPurple,

packages/core/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@google/gemini-cli-core",
-  "version": "0.39.0-nightly.20260408.e77b22e63",
+  "version": "0.40.0-nightly.20260414.g5b1f7375a",
   "description": "Gemini CLI Core",
   "license": "Apache-2.0",
   "repository": {
@@ -20,7 +20,8 @@
     "typecheck": "tsc --noEmit"
   },
   "files": [
-    "dist"
+    "dist",
+    "vendor"
   ],
   "dependencies": {
     "@a2a-js/sdk": "0.3.11",
@@ -31,7 +32,6 @@
     "@google/genai": "1.30.0",
     "@grpc/grpc-js": "^1.14.3",
     "@iarna/toml": "^2.2.5",
-    "@joshua.litt/get-ripgrep": "^0.0.3",
     "@modelcontextprotocol/sdk": "^1.23.0",
     "@opentelemetry/api": "^1.9.0",
     "@opentelemetry/api-logs": "^0.211.0",
@@ -59,6 +59,7 @@
     "diff": "^8.0.3",
     "dotenv": "^17.2.4",
     "dotenv-expand": "^12.0.3",
+    "execa": "^9.6.1",
     "fast-levenshtein": "^2.0.6",
     "fdir": "^6.4.6",
     "fzf": "^0.5.2",

packages/core/src/agents/skill-extraction-agent.test.ts

@@ -0,0 +1,90 @@
+/**
+ * @license
+ * Copyright 2026 Google LLC
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+import { describe, expect, it } from 'vitest';
+import { SkillExtractionAgent } from './skill-extraction-agent.js';
+import {
+  EDIT_TOOL_NAME,
+  GLOB_TOOL_NAME,
+  GREP_TOOL_NAME,
+  LS_TOOL_NAME,
+  READ_FILE_TOOL_NAME,
+  WRITE_FILE_TOOL_NAME,
+} from '../tools/tool-names.js';
+import { PREVIEW_GEMINI_FLASH_MODEL } from '../config/models.js';
+
+describe('SkillExtractionAgent', () => {
+  const skillsDir = '/tmp/skills';
+  const sessionIndex =
+    '[NEW] Debug login flow (12 user msgs) — /tmp/chats/session-1.json';
+  const existingSkillsSummary =
+    '## Workspace Skills (.gemini/skills — do NOT duplicate)\n- **existing-skill**: Existing description';
+
+  const agent = SkillExtractionAgent(
+    skillsDir,
+    sessionIndex,
+    existingSkillsSummary,
+  );
+
+  it('should expose expected metadata, model, and tools', () => {
+    expect(agent.kind).toBe('local');
+    expect(agent.name).toBe('confucius');
+    expect(agent.displayName).toBe('Skill Extractor');
+    expect(agent.modelConfig.model).toBe(PREVIEW_GEMINI_FLASH_MODEL);
+    expect(agent.toolConfig?.tools).toEqual(
+      expect.arrayContaining([
+        READ_FILE_TOOL_NAME,
+        WRITE_FILE_TOOL_NAME,
+        EDIT_TOOL_NAME,
+        LS_TOOL_NAME,
+        GLOB_TOOL_NAME,
+        GREP_TOOL_NAME,
+      ]),
+    );
+  });
+
+  it('should default to no skill unless recurrence and durability are proven', () => {
+    const prompt = agent.promptConfig.systemPrompt;
+
+    expect(prompt).toContain('Default to NO SKILL.');
+    expect(prompt).toContain(
+      'strong evidence this will recur for future agents in this repo/workflow',
+    );
+    expect(prompt).toContain('broader than a single incident');
+    expect(prompt).toContain('A skill MUST meet ALL of these criteria:');
+    expect(prompt).toContain(
+      'Future agents in this repo/workflow are likely to need it',
+    );
+  });
+
+  it('should explicitly reject one-off incidents and single-session preferences', () => {
+    const prompt = agent.promptConfig.systemPrompt;
+
+    expect(prompt).toContain('Single-session preferences');
+    expect(prompt).toContain('One-off incidents');
+    expect(prompt).toContain('Output-style preferences');
+    expect(prompt).toContain('cannot survive renaming the specific');
+  });
+
+  it('should warn that session summaries are user-intent summaries, not workflow evidence', () => {
+    const query = agent.promptConfig.query ?? '';
+
+    expect(query).toContain(existingSkillsSummary);
+    expect(query).toContain(sessionIndex);
+    expect(query).toContain(
+      'The summary is a user-intent summary, not a workflow summary.',
+    );
+    expect(query).toContain(
+      'The session summaries describe user intent, not workflow details.',
+    );
+    expect(query).toContain(
+      'Only write a skill if the evidence shows a durable, recurring workflow',
+    );
+    expect(query).toContain(
+      'If recurrence or future reuse is unclear, create no skill and explain why.',
+    );
+  });
+});

packages/core/src/agents/skill-extraction-agent.ts

@@ -7,11 +7,13 @@
 import { z } from 'zod';
 import type { LocalAgentDefinition } from './types.js';
 import {
+  ACTIVATE_SKILL_TOOL_NAME,
   EDIT_TOOL_NAME,
   GLOB_TOOL_NAME,
   GREP_TOOL_NAME,
   LS_TOOL_NAME,
   READ_FILE_TOOL_NAME,
+  SHELL_TOOL_NAME,
   WRITE_FILE_TOOL_NAME,
 } from '../tools/tool-names.js';
 import { PREVIEW_GEMINI_FLASH_MODEL } from '../config/models.js';
@@ -36,7 +38,7 @@ function buildSystemPrompt(skillsDir: string): string {
     '- solve similar tasks with fewer tool calls and fewer reasoning tokens',
     '- reuse proven workflows and verification checklists',
     '- avoid known failure modes and landmines',
-    '- anticipate user preferences without being reminded',
+    '- capture durable workflow constraints that future agents are likely to encounter again',
     '',
     '============================================================',
     'SAFETY AND HYGIENE (STRICT)',
@@ -59,6 +61,10 @@ function buildSystemPrompt(skillsDir: string): string {
     '1. "Is this something a competent agent would NOT already know?" If no, STOP.',
     '2. "Does an existing skill (listed below) already cover this?" If yes, STOP.',
     '3. "Can I write a concrete, step-by-step procedure?" If no, STOP.',
+    '4. "Is there strong evidence this will recur for future agents in this repo/workflow?" If no, STOP.',
+    '5. "Is this broader than a single incident (one bug, one ticket, one branch, one date, one exact error)?" If no, STOP.',
+    '',
+    'Default to NO SKILL.',
     '',
     'Do NOT create skills for:',
     '',
@@ -67,6 +73,10 @@ function buildSystemPrompt(skillsDir: string): string {
     '- **Pure Q&A**: The user asked "how does X work?" and got an answer. No procedure.',
     '- **Brainstorming/design**: Discussion of how to build something, without a validated',
     '  implementation that produced a reusable procedure.',
+    '- **Single-session preferences**: User-specific style/output preferences or workflow',
+    '  preferences mentioned only once.',
+    '- **One-off incidents**: Debugging or incident response tied to a single bug, ticket,',
+    '  branch, date, or exact error string.',
     '- **Anything already covered by an existing skill** (global, workspace, builtin, or',
     '  previously extracted). Check the "Existing Skills" section carefully.',
     '',
@@ -74,31 +84,40 @@ function buildSystemPrompt(skillsDir: string): string {
     'WHAT COUNTS AS A SKILL',
     '============================================================',
     '',
-    'A skill MUST meet BOTH of these criteria:',
+    'A skill MUST meet ALL of these criteria:',
     '',
     '1. **Procedural and concrete**: It can be expressed as numbered steps with specific',
     '   commands, paths, or code patterns. If you can only write vague guidance, it is NOT',
     '   a skill. "Be careful with X" is advice, not a skill.',
     '',
-    '2. **Non-obvious and project-specific**: A competent agent would NOT already know this.',
-    '   It encodes project-specific knowledge, non-obvious ordering constraints, or',
-    '   hard-won failure shields that cannot be inferred from the codebase alone.',
+    '2. **Durable and reusable**: Future agents in this repo/workflow are likely to need it',
+    '   again. If it only solved one incident, it is NOT a skill.',
     '',
-    'Confidence tiers (prefer higher tiers):',
+    '3. **Evidence-backed and project-specific**: It encodes project-specific knowledge,',
+    '   repeated operational constraints, or hard-won failure shields supported by session',
+    '   evidence. Do not assume something is non-obvious just because it sounds detailed.',
     '',
-    '**High confidence** — create the skill:',
-    '- The same workflow appeared in multiple sessions (cross-session repetition)',
-    '- A multi-step procedure was validated (tests passed, user confirmed success)',
+    'Confidence tiers:',
     '',
-    '**Medium confidence** — create the skill if it is clearly project-specific:',
-    '- A project-specific build/test/deploy/release procedure was established',
-    '- A non-obvious ordering constraint or prerequisite was discovered',
-    '- A failure mode was hit and a concrete fix was found and verified',
+    '**High confidence** — create the skill only when recurrence/durability is clear:',
+    '- The same workflow appeared in multiple sessions (cross-session repetition), OR it is',
+    '  a stable recurring repo workflow (for example setup/build/test/deploy/release) with a',
+    '  clear future trigger',
+    '- The workflow was validated (tests passed, user confirmed success, or the same fix',
+    '  worked repeatedly)',
+    '- The skill can be named without referencing a specific incident, bug, branch, or date',
+    '',
+    '**Medium confidence** — usually do NOT create the skill yet:',
+    '- A project-specific procedure appeared once and seems useful, but recurrence is not yet',
+    '  clear',
+    '- A verified fix exists, but it is still tied to one incident',
+    '- A user correction changed the approach once, but durability is uncertain',
     '',
     '**Low confidence** — do NOT create the skill:',
     '- A one-off debugging session with no reusable procedure',
     '- Generic workflows any agent could figure out from the codebase',
     '- A code review or investigation with no durable takeaway',
+    '- Output-style preferences that do not materially change procedure',
     '',
     'Aim for 0-2 skills per run. Quality over quantity.',
     '',
@@ -117,8 +136,10 @@ function buildSystemPrompt(skillsDir: string): string {
     '',
     'What to look for:',
     '',
-    '- User corrections: "No, do it this way" -> preference signal',
+    '- User corrections that change procedure in a durable way, especially when repeated',
+    '  across sessions',
     '- Repeated patterns across sessions: same commands, same file paths, same workflow',
+    '- Stable recurring repo lifecycle workflows with clear future triggers',
     '- Failed attempts followed by successful ones -> failure shield',
     '- Multi-step procedures that were validated (tests passed, user confirmed)',
     '- User interruptions: "Stop, you need to X first" -> ordering constraint',
@@ -129,45 +150,8 @@ function buildSystemPrompt(skillsDir: string): string {
     '- Tool outputs that are just data (file contents, search results)',
     '- Speculative plans that were never executed',
     "- Temporary context (current branch name, today's date, specific error IDs)",
-    '',
-    '============================================================',
-    'SKILL FORMAT',
-    '============================================================',
-    '',
-    'Each skill is a directory containing a SKILL.md file with YAML frontmatter',
-    'and optional supporting scripts.',
-    '',
-    'Directory structure:',
-    `  ${skillsDir}/<skill-name>/`,
-    '    SKILL.md            # Required entrypoint',
-    '    scripts/<tool>.*    # Optional helper scripts (Python stdlib-only or shell)',
-    '',
-    'SKILL.md structure:',
-    '',
-    '  ---',
-    '  name: <skill-name>',
-    '  description: <1-2 lines; include concrete triggers in user-like language>',
-    '  ---',
-    '',
-    '  ## When to Use',
-    '  <Clear trigger conditions and non-goals>',
-    '',
-    '  ## Procedure',
-    '  <Numbered steps with specific commands, paths, code patterns>',
-    '',
-    '  ## Pitfalls and Fixes',
-    '  <symptom -> likely cause -> fix; only include observed failures>',
-    '',
-    '  ## Verification',
-    '  <Concrete success checks>',
-    '',
-    'Supporting scripts (optional but recommended when applicable):',
-    '- Put helper scripts in scripts/ and reference them from SKILL.md',
-    '- Prefer Python (stdlib only) or small shell scripts',
-    '- Make scripts safe: no destructive actions, no secrets, deterministic output',
-    '- Include a usage example in SKILL.md',
-    '',
-    'Naming: kebab-case (e.g., fix-lint-errors, run-migrations).',
+    '- Similar session summaries without matching workflow evidence',
+    '- One-off artifact names: bug IDs, branch names, timestamps, exact incident strings',
     '',
     '============================================================',
     'UPDATING EXISTING SKILLS (PATCHES)',
@@ -214,7 +198,10 @@ function buildSystemPrompt(skillsDir: string): string {
     '- Keep scopes distinct. Avoid overlapping "do-everything" skills.',
     '- Every skill MUST have: triggers, procedure, at least one pitfall or verification step.',
     '- If you cannot write a reliable procedure (too many unknowns), do NOT create the skill.',
-    '- Do not create skills for generic advice that any competent agent would already know.',
+    '- If the candidate is tied to one incident or cannot survive renaming the specific',
+    '  bug/ticket, do NOT create it.',
+    '- Do not create skills for generic advice, output-style preferences, or ephemeral',
+    '  choices that any competent agent would already know or adapt to on the fly.',
     '- Prefer fewer, higher-quality skills. 0-2 skills per run is typical. 3+ is unusual.',
     '',
     '============================================================',
@@ -223,18 +210,32 @@ function buildSystemPrompt(skillsDir: string): string {
     '',
     `1. Use list_directory on ${skillsDir} to see existing skills.`,
     '2. If skills exist, read their SKILL.md files to understand what is already captured.',
-    '3. Scan the session index provided in the query. Look for [NEW] sessions whose summaries',
-    '   suggest workflows that ALSO appear in other sessions (either [NEW] or [old]).',
-    '4. Apply the minimum signal gate. If no repeated patterns are visible, report that and finish.',
-    '5. For promising patterns, use read_file on the session file paths to inspect the full',
-    '   conversation. Confirm the workflow was actually repeated and validated.',
-    '6. For each confirmed skill, verify it meets ALL criteria (repeatable, procedural, high-leverage).',
-    '7. Write new SKILL.md files or update existing ones in your directory using write_file.',
+    '3. Use activate_skill to load the "skill-creator" skill. Follow its design guidance',
+    '   (conciseness, progressive disclosure, frontmatter format, bundled resources) when',
+    '   writing SKILL.md files. You may also use its init_skill.cjs script to scaffold new',
+    '   skill directories and package_skill.cjs to validate finished skills.',
+    '   IMPORTANT: You are a background agent with no user interaction. Skip any interactive',
+    '   steps in the skill-creator guide (asking clarifying questions, requesting user feedback,',
+    '   installation prompts, iteration loops). Use only its format and quality guidance.',
+    '4. Scan the session index provided in the query. Look for [NEW] sessions whose summaries',
+    '   hint at workflows that ALSO appear in other sessions (either [NEW] or [old]) or at a',
+    '   stable recurring repo workflow. Remember: summary similarity alone is NOT enough.',
+    '5. Apply the minimum signal gate. If recurrence or durability is not visible, report that',
+    '   no skill should be created and finish.',
+    '6. For promising patterns, use read_file on the session file paths to inspect the full',
+    '   conversation. Confirm the workflow was actually repeated and validated. Read at least',
+    '   two sessions unless the candidate is clearly a stable recurring repo lifecycle workflow.',
+    '7. For each candidate, verify it meets ALL criteria. Before writing, make sure you can',
+    '   state: future trigger, evidence sessions, recurrence signal, validation signal, and',
+    '   why it is not generic.',
+    '8. Write new SKILL.md files or update existing ones in your directory.',
+    '   Use run_shell_command to run init_skill.cjs for scaffolding and package_skill.cjs for validation.',
     '   For skills that live OUTSIDE your directory, write a .patch file instead (see UPDATING EXISTING SKILLS).',
-    '8. Write COMPLETE files — never partially update a SKILL.md.',
+    '9. Write COMPLETE files — never partially update a SKILL.md.',
     '',
     'IMPORTANT: Do NOT read every session. Only read sessions whose summaries suggest a',
-    'repeated pattern worth investigating. Most runs should read 0-3 sessions and create 0 skills.',
+    'repeated pattern or a stable recurring repo workflow worth investigating. Most runs',
+    'should read 0-3 sessions and create 0 skills.',
     'Do not explore the codebase. Work only with the session index, session files, and the skills directory.',
   ].join('\n');
 }
@@ -244,8 +245,9 @@ function buildSystemPrompt(skillsDir: string): string {
  * writes reusable SKILL.md files to the project memory directory.
  *
  * This agent is designed to run in the background on session startup.
- * It has restricted tool access (file tools only, no shell or user interaction)
- * and is prompted to only operate within the skills memory directory.
+ * It has restricted tool access (file tools, shell, and skill activation — no
+ * user interaction) and is prompted to only operate within the skills memory
+ * directory.
  */
 export const SkillExtractionAgent = (
   skillsDir: string,
@@ -279,12 +281,14 @@ export const SkillExtractionAgent = (
   },
   toolConfig: {
     tools: [
+      ACTIVATE_SKILL_TOOL_NAME,
       READ_FILE_TOOL_NAME,
       WRITE_FILE_TOOL_NAME,
       EDIT_TOOL_NAME,
       LS_TOOL_NAME,
       GLOB_TOOL_NAME,
       GREP_TOOL_NAME,
+      SHELL_TOOL_NAME,
     ],
   },
   get promptConfig() {
@@ -301,6 +305,9 @@ export const SkillExtractionAgent = (
         'Below is an index of past conversation sessions. Each line shows:',
         '[NEW] or [old] status, a 1-line summary, message count, and the file path.',
         '',
+        'The summary is a user-intent summary, not a workflow summary.',
+        'Matching summary text alone is never enough evidence for a reusable skill.',
+        '',
         '[NEW] = not yet processed for skill extraction (focus on these)',
         '[old] = previously processed (read only if a [NEW] session hints at a repeated pattern)',
         '',
@@ -319,7 +326,7 @@ export const SkillExtractionAgent = (
 
     return {
       systemPrompt: buildSystemPrompt(skillsDir),
-      query: `${initialContext}\n\nAnalyze the session index above. Read sessions that suggest repeated workflows using read_file. Extract reusable skills to ${skillsDir}/.`,
+      query: `${initialContext}\n\nAnalyze the session index above. The session summaries describe user intent, not workflow details. Read sessions that suggest repeated workflows using read_file. Only write a skill if the evidence shows a durable, recurring workflow or a stable recurring repo procedure. If recurrence or future reuse is unclear, create no skill and explain why.`,
     };
   },
   runConfig: {

packages/core/src/config/config.ts

@@ -3552,6 +3552,7 @@ export class Config implements McpContext, AgentLoopContext {
           registry.registerTool(new RipGrepTool(this, this.messageBus)),
         );
       } else {
+        debugLogger.warn(`Ripgrep is not available. Falling back to GrepTool.`);
         logRipgrepFallback(this, new RipgrepFallbackEvent(errorString));
         maybeRegister(GrepTool, () =>
           registry.registerTool(new GrepTool(this, this.messageBus)),

packages/core/src/context/toolOutputMaskingService.test.ts

@@ -661,4 +661,82 @@ describe('ToolOutputMaskingService', () => {
       )['output'],
     ).toContain(MASKING_INDICATOR_TAG);
   });
+
+  it('should use existing outputFile if available in the tool response', async () => {
+    // Setup: Create a large history to trigger masking
+    const largeContent = 'a'.repeat(60000);
+    const existingOutputFile = path.join(testTempDir, 'truly_full_output.txt');
+    await fs.promises.writeFile(existingOutputFile, 'truly full content');
+
+    const history: Content[] = [
+      {
+        role: 'user',
+        parts: [{ text: 'Old turn' }],
+      },
+      {
+        role: 'model',
+        parts: [
+          {
+            functionResponse: {
+              name: 'shell',
+              id: 'call-1',
+              response: {
+                output: largeContent,
+                outputFile: existingOutputFile,
+              },
+            },
+          },
+        ],
+      },
+      // Protection buffer
+      {
+        role: 'user',
+        parts: [
+          {
+            functionResponse: {
+              name: 'padding',
+              response: { output: 'B'.repeat(60000) },
+            },
+          },
+        ],
+      },
+      {
+        role: 'user',
+        parts: [{ text: 'Newest turn' }],
+      },
+    ];
+
+    mockedEstimateTokenCountSync.mockImplementation((parts: Part[]) => {
+      const resp = parts[0].functionResponse?.response as Record<
+        string,
+        unknown
+      >;
+      const content = (resp?.['output'] as string) ?? JSON.stringify(resp);
+      if (content.includes(`<${MASKING_INDICATOR_TAG}`)) return 100;
+
+      const name = parts[0].functionResponse?.name;
+      if (name === 'shell') return 60000;
+      if (name === 'padding') return 60000;
+      return 10;
+    });
+
+    // Trigger masking
+    const result = await service.mask(history, mockConfig);
+
+    expect(result.maskedCount).toBe(2);
+    const maskedPart = result.newHistory[1].parts![0];
+    const maskedResponse = maskedPart.functionResponse?.response as Record<
+      string,
+      unknown
+    >;
+    const maskedOutput = maskedResponse['output'] as string;
+
+    // Verify the masked snippet points to the existing file
+    expect(maskedOutput).toContain(
+      `Full output available at: ${existingOutputFile}`,
+    );
+
+    // Verify the path in maskedOutput is exactly the one we provided
+    expect(maskedOutput).toContain(existingOutputFile);
+  });
 });

packages/core/src/context/toolOutputMaskingService.ts

@@ -48,6 +48,18 @@ export interface MaskingResult {
   tokensSaved: number;
 }
 
+interface HasOutputFile {
+  outputFile: string;
+}
+
+function hasOutputFile(obj: unknown): obj is HasOutputFile {
+  if (typeof obj !== 'object' || obj === null || !('outputFile' in obj)) {
+    return false;
+  }
+  const val = (obj as Record<string, unknown>)['outputFile'];
+  return typeof val === 'string';
+}
+
 /**
  * Service to manage context window efficiency by masking bulky tool outputs (Tool Output Masking).
  *
@@ -182,25 +194,44 @@ export class ToolOutputMaskingService {
 
       const toolName = part.functionResponse.name || 'unknown_tool';
       const callId = part.functionResponse.id || Date.now().toString();
-      const safeToolName = sanitizeFilenamePart(toolName).toLowerCase();
-      const safeCallId = sanitizeFilenamePart(callId).toLowerCase();
-      const fileName = `${safeToolName}_${safeCallId}_${Math.random()
-        .toString(36)
-        .substring(7)}.txt`;
-      const filePath = path.join(toolOutputsDir, fileName);
-
-      await fsPromises.writeFile(filePath, content, 'utf-8');
 
       const originalResponse =
         // eslint-disable-next-line @typescript-eslint/no-unsafe-type-assertion
         (part.functionResponse.response as Record<string, unknown>) || {};
 
-      const totalLines = content.split('\n').length;
-      const fileSizeMB = (
-        Buffer.byteLength(content, 'utf8') /
-        1024 /
-        1024
-      ).toFixed(2);
+      let filePath = '';
+      let fileSizeMB = '0.00';
+      let totalLines = 0;
+
+      if (hasOutputFile(originalResponse) && originalResponse.outputFile) {
+        filePath = originalResponse.outputFile;
+        try {
+          const stats = await fsPromises.stat(filePath);
+          fileSizeMB = (stats.size / 1024 / 1024).toFixed(2);
+          // For truly full files, we don't count lines as it's too slow.
+          // We just indicate it's the full file.
+          totalLines = -1;
+        } catch {
+          // Fallback if file is gone
+          filePath = '';
+        }
+      }
+
+      if (!filePath) {
+        const safeToolName = sanitizeFilenamePart(toolName).toLowerCase();
+        const safeCallId = sanitizeFilenamePart(callId).toLowerCase();
+        const fileName = `${safeToolName}_${safeCallId}_${Math.random()
+          .toString(36)
+          .substring(7)}.txt`;
+        filePath = path.join(toolOutputsDir, fileName);
+
+        await fsPromises.writeFile(filePath, content, 'utf-8');
+
+        totalLines = content.split('\n').length;
+        fileSizeMB = (Buffer.byteLength(content, 'utf8') / 1024 / 1024).toFixed(
+          2,
+        );
+      }
 
       let preview = '';
       if (toolName === SHELL_TOOL_NAME) {

packages/core/src/core/contentGenerator.test.ts

@@ -148,7 +148,7 @@ describe('createContentGenerator', () => {
     );
     expect(GoogleGenAI).toHaveBeenCalledWith({
       apiKey: 'test-api-key',
-      vertexai: undefined,
+      vertexai: false,
       httpOptions: expect.objectContaining({
         headers: expect.objectContaining({
           'User-Agent': expect.stringMatching(
@@ -365,7 +365,7 @@ describe('createContentGenerator', () => {
 
     expect(GoogleGenAI).toHaveBeenCalledWith({
       apiKey: 'test-api-key',
-      vertexai: undefined,
+      vertexai: false,
       httpOptions: expect.objectContaining({
         headers: expect.objectContaining({
           'User-Agent': expect.any(String),
@@ -409,7 +409,7 @@ describe('createContentGenerator', () => {
 
     expect(GoogleGenAI).toHaveBeenCalledWith({
       apiKey: 'test-api-key',
-      vertexai: undefined,
+      vertexai: false,
       httpOptions: expect.objectContaining({
         headers: expect.objectContaining({
           'User-Agent': expect.any(String),
@@ -443,7 +443,7 @@ describe('createContentGenerator', () => {
 
     expect(GoogleGenAI).toHaveBeenCalledWith({
       apiKey: 'test-api-key',
-      vertexai: undefined,
+      vertexai: false,
       httpOptions: expect.objectContaining({
         headers: expect.objectContaining({
           'User-Agent': expect.any(String),
@@ -481,7 +481,7 @@ describe('createContentGenerator', () => {
     );
     expect(GoogleGenAI).toHaveBeenCalledWith({
       apiKey: 'test-api-key',
-      vertexai: undefined,
+      vertexai: false,
       httpOptions: expect.objectContaining({
         headers: {
           'User-Agent': expect.any(String),
@@ -517,7 +517,7 @@ describe('createContentGenerator', () => {
 
     expect(GoogleGenAI).toHaveBeenCalledWith({
       apiKey: 'test-api-key',
-      vertexai: undefined,
+      vertexai: false,
       httpOptions: expect.objectContaining({
         headers: expect.objectContaining({
           'User-Agent': expect.any(String),
@@ -550,7 +550,7 @@ describe('createContentGenerator', () => {
 
     expect(GoogleGenAI).toHaveBeenCalledWith({
       apiKey: 'test-api-key',
-      vertexai: undefined,
+      vertexai: false,
       httpOptions: expect.objectContaining({
         headers: expect.objectContaining({
           'User-Agent': expect.any(String),
@@ -589,7 +589,7 @@ describe('createContentGenerator', () => {
 
     expect(GoogleGenAI).toHaveBeenCalledWith({
       apiKey: 'test-api-key',
-      vertexai: undefined,
+      vertexai: false,
       httpOptions: expect.objectContaining({
         headers: expect.objectContaining({
           'User-Agent': expect.any(String),
@@ -638,6 +638,193 @@ describe('createContentGenerator', () => {
       apiVersion: 'v1alpha',
     });
   });
+
+  it('should pass baseUrl to GoogleGenAI when GOOGLE_GEMINI_BASE_URL is set', async () => {
+    const mockConfig = {
+      getModel: vi.fn().mockReturnValue('gemini-pro'),
+      getProxy: vi.fn().mockReturnValue(undefined),
+      getUsageStatisticsEnabled: () => false,
+      getClientName: vi.fn().mockReturnValue(undefined),
+    } as unknown as Config;
+
+    const mockGenerator = {
+      models: {},
+    } as unknown as GoogleGenAI;
+    vi.mocked(GoogleGenAI).mockImplementation(() => mockGenerator as never);
+    vi.stubEnv('GOOGLE_GEMINI_BASE_URL', 'https://gemini.test.local');
+    vi.stubEnv('GEMINI_API_KEY', 'test-api-key');
+
+    const config = await createContentGeneratorConfig(
+      mockConfig,
+      AuthType.USE_GEMINI,
+    );
+    await createContentGenerator(config, mockConfig);
+
+    expect(GoogleGenAI).toHaveBeenCalledWith(
+      expect.objectContaining({
+        apiKey: 'test-api-key',
+        vertexai: false,
+        httpOptions: expect.objectContaining({
+          baseUrl: 'https://gemini.test.local',
+        }),
+      }),
+    );
+  });
+
+  it('should pass baseUrl to GoogleGenAI when GOOGLE_VERTEX_BASE_URL is set', async () => {
+    const mockConfig = {
+      getModel: vi.fn().mockReturnValue('gemini-pro'),
+      getProxy: vi.fn().mockReturnValue(undefined),
+      getUsageStatisticsEnabled: () => false,
+      getClientName: vi.fn().mockReturnValue(undefined),
+    } as unknown as Config;
+
+    const mockGenerator = {
+      models: {},
+    } as unknown as GoogleGenAI;
+    vi.mocked(GoogleGenAI).mockImplementation(() => mockGenerator as never);
+    vi.stubEnv('GOOGLE_VERTEX_BASE_URL', 'https://vertex.test.local');
+    vi.stubEnv('GOOGLE_CLOUD_PROJECT', 'my-project');
+    vi.stubEnv('GOOGLE_CLOUD_LOCATION', 'us-central1');
+
+    const config = await createContentGeneratorConfig(
+      mockConfig,
+      AuthType.USE_VERTEX_AI,
+    );
+    await createContentGenerator(config, mockConfig);
+
+    expect(GoogleGenAI).toHaveBeenCalledWith(
+      expect.objectContaining({
+        apiKey: undefined,
+        vertexai: true,
+        httpOptions: expect.objectContaining({
+          baseUrl: 'https://vertex.test.local',
+        }),
+      }),
+    );
+  });
+
+  it('should prefer GOOGLE_VERTEX_BASE_URL when authType is USE_VERTEX_AI without inferred vertex credentials', async () => {
+    const mockConfig = {
+      getModel: vi.fn().mockReturnValue('gemini-pro'),
+      getProxy: vi.fn().mockReturnValue(undefined),
+      getUsageStatisticsEnabled: () => false,
+      getClientName: vi.fn().mockReturnValue(undefined),
+    } as unknown as Config;
+
+    const mockGenerator = {
+      models: {},
+    } as unknown as GoogleGenAI;
+    vi.mocked(GoogleGenAI).mockImplementation(() => mockGenerator as never);
+    vi.stubEnv('GOOGLE_GEMINI_BASE_URL', 'https://gemini.test.local');
+    vi.stubEnv('GOOGLE_VERTEX_BASE_URL', 'https://vertex.test.local');
+
+    await createContentGenerator(
+      {
+        authType: AuthType.USE_VERTEX_AI,
+      },
+      mockConfig,
+    );
+
+    expect(GoogleGenAI).toHaveBeenCalledWith(
+      expect.objectContaining({
+        apiKey: undefined,
+        vertexai: true,
+        httpOptions: expect.objectContaining({
+          baseUrl: 'https://vertex.test.local',
+        }),
+      }),
+    );
+  });
+
+  it('should prefer an explicit baseUrl over GOOGLE_GEMINI_BASE_URL', async () => {
+    const mockConfig = {
+      getModel: vi.fn().mockReturnValue('gemini-pro'),
+      getProxy: vi.fn().mockReturnValue(undefined),
+      getUsageStatisticsEnabled: () => false,
+      getClientName: vi.fn().mockReturnValue(undefined),
+    } as unknown as Config;
+
+    const mockGenerator = {
+      models: {},
+    } as unknown as GoogleGenAI;
+    vi.mocked(GoogleGenAI).mockImplementation(() => mockGenerator as never);
+    vi.stubEnv('GOOGLE_GEMINI_BASE_URL', 'https://env.test.local');
+    vi.stubEnv('GEMINI_API_KEY', 'test-api-key');
+
+    const config = await createContentGeneratorConfig(
+      mockConfig,
+      AuthType.USE_GEMINI,
+      undefined,
+      'https://explicit.test.local',
+    );
+    await createContentGenerator(config, mockConfig);
+
+    expect(GoogleGenAI).toHaveBeenCalledWith(
+      expect.objectContaining({
+        httpOptions: expect.objectContaining({
+          baseUrl: 'https://explicit.test.local',
+        }),
+      }),
+    );
+  });
+
+  it('should allow localhost baseUrl overrides over http', async () => {
+    const mockConfig = {
+      getModel: vi.fn().mockReturnValue('gemini-pro'),
+      getProxy: vi.fn().mockReturnValue(undefined),
+      getUsageStatisticsEnabled: () => false,
+      getClientName: vi.fn().mockReturnValue(undefined),
+    } as unknown as Config;
+
+    const mockGenerator = {
+      models: {},
+    } as unknown as GoogleGenAI;
+    vi.mocked(GoogleGenAI).mockImplementation(() => mockGenerator as never);
+
+    await createContentGenerator(
+      {
+        apiKey: 'test-api-key',
+        authType: AuthType.USE_GEMINI,
+        baseUrl: 'http://127.0.0.1:8080',
+      },
+      mockConfig,
+    );
+
+    expect(GoogleGenAI).toHaveBeenCalledWith(
+      expect.objectContaining({
+        httpOptions: expect.objectContaining({
+          baseUrl: 'http://127.0.0.1:8080',
+        }),
+      }),
+    );
+  });
+
+  it('should reject invalid custom baseUrl values', async () => {
+    await expect(
+      createContentGenerator(
+        {
+          apiKey: 'test-api-key',
+          authType: AuthType.USE_GEMINI,
+          baseUrl: 'not-a-url',
+        },
+        mockConfig,
+      ),
+    ).rejects.toThrow('Invalid custom base URL: not-a-url');
+  });
+
+  it('should reject non-https remote custom baseUrl values', async () => {
+    await expect(
+      createContentGenerator(
+        {
+          apiKey: 'test-api-key',
+          authType: AuthType.USE_GEMINI,
+          baseUrl: 'http://example.com',
+        },
+        mockConfig,
+      ),
+    ).rejects.toThrow('Custom base URL must use HTTPS unless it is localhost.');
+  });
 });
 
 describe('createContentGeneratorConfig', () => {

packages/core/src/core/contentGenerator.ts

@@ -101,6 +101,21 @@ export type ContentGeneratorConfig = {
   customHeaders?: Record<string, string>;
 };
 
+const LOCAL_HOSTNAMES = ['localhost', '127.0.0.1', '[::1]'];
+
+function validateBaseUrl(baseUrl: string): void {
+  let url: URL;
+  try {
+    url = new URL(baseUrl);
+  } catch {
+    throw new Error(`Invalid custom base URL: ${baseUrl}`);
+  }
+
+  if (url.protocol !== 'https:' && !LOCAL_HOSTNAMES.includes(url.hostname)) {
+    throw new Error('Custom base URL must use HTTPS unless it is localhost.');
+  }
+}
+
 export async function createContentGeneratorConfig(
   config: Config,
   authType: AuthType | undefined,
@@ -273,18 +288,32 @@ export async function createContentGenerator(
           'x-gemini-api-privileged-user-id': `${installationId}`,
         };
       }
+      let baseUrl = config.baseUrl;
+      if (!baseUrl) {
+        const envBaseUrl =
+          config.authType === AuthType.USE_VERTEX_AI
+            ? process.env['GOOGLE_VERTEX_BASE_URL']
+            : process.env['GOOGLE_GEMINI_BASE_URL'];
+        if (envBaseUrl) {
+          validateBaseUrl(envBaseUrl);
+          baseUrl = envBaseUrl;
+        }
+      } else {
+        validateBaseUrl(baseUrl);
+      }
+
       const httpOptions: {
         baseUrl?: string;
         headers: Record<string, string>;
       } = { headers };
 
-      if (config.baseUrl) {
-        httpOptions.baseUrl = config.baseUrl;
+      if (baseUrl) {
+        httpOptions.baseUrl = baseUrl;
       }
 
       const googleGenAI = new GoogleGenAI({
         apiKey: config.apiKey === '' ? undefined : config.apiKey,
-        vertexai: config.vertexai,
+        vertexai: config.vertexai ?? config.authType === AuthType.USE_VERTEX_AI,
         httpOptions,
         ...(apiVersionEnv && { apiVersion: apiVersionEnv }),
       });

packages/core/src/core/geminiChat_network_retry.test.ts

@@ -519,4 +519,70 @@ describe('GeminiChat Network Retries', () => {
       }),
     );
   });
+
+  it('should retry on OpenSSL 3.x SSL error during stream iteration (ERR_SSL_SSL/TLS_ALERT_BAD_RECORD_MAC)', async () => {
+    // OpenSSL 3.x produces a different error code format than OpenSSL 1.x
+    const sslError = new Error(
+      'request to https://cloudcode-pa.googleapis.com/v1internal:streamGenerateContent failed',
+    ) as NodeJS.ErrnoException & { type?: string };
+    sslError.type = 'system';
+    sslError.errno =
+      'ERR_SSL_SSL/TLS_ALERT_BAD_RECORD_MAC' as unknown as number;
+    sslError.code = 'ERR_SSL_SSL/TLS_ALERT_BAD_RECORD_MAC';
+
+    vi.mocked(mockContentGenerator.generateContentStream)
+      .mockImplementationOnce(async () =>
+        (async function* () {
+          yield {
+            candidates: [
+              { content: { parts: [{ text: 'Partial response...' }] } },
+            ],
+          } as unknown as GenerateContentResponse;
+          throw sslError;
+        })(),
+      )
+      .mockImplementationOnce(async () =>
+        (async function* () {
+          yield {
+            candidates: [
+              {
+                content: { parts: [{ text: 'Complete response after retry' }] },
+                finishReason: 'STOP',
+              },
+            ],
+          } as unknown as GenerateContentResponse;
+        })(),
+      );
+
+    const stream = await chat.sendMessageStream(
+      { model: 'test-model' },
+      'test message',
+      'prompt-id-ssl3-mid-stream',
+      new AbortController().signal,
+      LlmRole.MAIN,
+    );
+
+    const events: StreamEvent[] = [];
+    for await (const event of stream) {
+      events.push(event);
+    }
+
+    const retryEvent = events.find((e) => e.type === StreamEventType.RETRY);
+    expect(retryEvent).toBeDefined();
+
+    const successChunk = events.find(
+      (e) =>
+        e.type === StreamEventType.CHUNK &&
+        e.value.candidates?.[0]?.content?.parts?.[0]?.text ===
+          'Complete response after retry',
+    );
+    expect(successChunk).toBeDefined();
+
+    expect(mockLogNetworkRetryAttempt).toHaveBeenCalledWith(
+      expect.anything(),
+      expect.objectContaining({
+        error_type: 'ERR_SSL_SSL/TLS_ALERT_BAD_RECORD_MAC',
+      }),
+    );
+  });
 });

packages/core/src/policy/policy-engine.test.ts

@@ -1762,6 +1762,39 @@ describe('PolicyEngine', () => {
   });
 
   describe('shell command parsing failure', () => {
+    it('should return ALLOW in YOLO mode for dangerous commands due to heuristics override', async () => {
+      // Create an engine with YOLO mode and a sandbox manager that flags a command as dangerous
+      const rules: PolicyRule[] = [
+        {
+          toolName: '*',
+          decision: PolicyDecision.ALLOW,
+          priority: 999,
+          modes: [ApprovalMode.YOLO],
+        },
+      ];
+
+      const mockSandboxManager = new NoopSandboxManager();
+      mockSandboxManager.isDangerousCommand = vi.fn().mockReturnValue(true);
+      mockSandboxManager.isKnownSafeCommand = vi.fn().mockReturnValue(false);
+
+      engine = new PolicyEngine({
+        rules,
+        approvalMode: ApprovalMode.YOLO,
+        sandboxManager: mockSandboxManager,
+      });
+
+      const result = await engine.check(
+        {
+          name: 'run_shell_command',
+          args: { command: 'powershell echo "dangerous"' },
+        },
+        undefined,
+      );
+
+      // Even though the command is flagged as dangerous, YOLO mode should preserve the ALLOW decision
+      expect(result.decision).toBe(PolicyDecision.ALLOW);
+    });
+
     it('should return ALLOW in YOLO mode even if shell command parsing fails', async () => {
       const { splitCommands } = await import('../utils/shell-utils.js');
       const rules: PolicyRule[] = [

packages/core/src/policy/policy-engine.ts

@@ -312,6 +312,13 @@ export class PolicyEngine {
       const parsedArgs = parsedObjArgs.map(extractStringFromParseEntry);
 
       if (this.sandboxManager.isDangerousCommand(parsedArgs)) {
+        if (this.approvalMode === ApprovalMode.YOLO) {
+          debugLogger.debug(
+            `[PolicyEngine.check] Command evaluated as dangerous, but YOLO mode is active. Preserving decision: ${command}`,
+          );
+          return decision;
+        }
+
         debugLogger.debug(
           `[PolicyEngine.check] Command evaluated as dangerous, forcing ASK_USER: ${command}`,
         );

packages/core/src/scheduler/tool-executor.test.ts

@@ -5,6 +5,7 @@
  */
 
 import { describe, it, expect, vi, beforeEach, afterEach } from 'vitest';
+import fsPromises from 'node:fs/promises';
 import { ToolExecutor } from './tool-executor.js';
 import {
   type Config,
@@ -32,6 +33,7 @@ import {
 vi.mock('../utils/fileUtils.js', () => ({
   saveTruncatedToolOutput: vi.fn(),
   formatTruncatedToolOutput: vi.fn(),
+  moveToolOutputToFile: vi.fn(),
 }));
 
 // Mock executeToolWithHooks
@@ -436,73 +438,10 @@ describe('ToolExecutor', () => {
       const response = result.response.responseParts[0]?.functionResponse
         ?.response as Record<string, unknown>;
       // The content should be the *truncated* version returned by the mock formatTruncatedToolOutput
-      expect(response).toEqual({ output: 'TruncatedContent...' });
-      expect(result.response.outputFile).toBe('/tmp/truncated_output.txt');
-    }
-  });
-
-  it('should truncate large MCP tool output with single text Part', async () => {
-    // 1. Setup Config for Truncation
-    vi.spyOn(config, 'getTruncateToolOutputThreshold').mockReturnValue(10);
-    vi.spyOn(config.storage, 'getProjectTempDir').mockReturnValue('/tmp');
-
-    const mcpToolName = 'get_big_text';
-    const messageBus = createMockMessageBus();
-    const mcpTool = new DiscoveredMCPTool(
-      {} as CallableTool,
-      'my-server',
-      'get_big_text',
-      'A test MCP tool',
-      {},
-      messageBus,
-    );
-    const invocation = mcpTool.build({});
-    const longText = 'This is a very long MCP output that should be truncated.';
-
-    // 2. Mock execution returning Part[] with single text Part
-    vi.mocked(coreToolHookTriggers.executeToolWithHooks).mockResolvedValue({
-      llmContent: [{ text: longText }],
-      returnDisplay: longText,
-    });
-
-    const scheduledCall: ScheduledToolCall = {
-      status: CoreToolCallStatus.Scheduled,
-      request: {
-        callId: 'call-mcp-trunc',
-        name: mcpToolName,
-        args: { query: 'test' },
-        isClientInitiated: false,
-        prompt_id: 'prompt-mcp-trunc',
-      },
-      tool: mcpTool,
-      invocation: invocation as unknown as AnyToolInvocation,
-      startTime: Date.now(),
-    };
-
-    // 3. Execute
-    const result = await executor.execute({
-      call: scheduledCall,
-      signal: new AbortController().signal,
-      onUpdateToolCall: vi.fn(),
-    });
-
-    // 4. Verify Truncation Logic
-    expect(fileUtils.saveTruncatedToolOutput).toHaveBeenCalledWith(
-      longText,
-      mcpToolName,
-      'call-mcp-trunc',
-      expect.any(String),
-      'test-session-id',
-    );
-
-    expect(fileUtils.formatTruncatedToolOutput).toHaveBeenCalledWith(
-      longText,
-      '/tmp/truncated_output.txt',
-      10,
-    );
-
-    expect(result.status).toBe(CoreToolCallStatus.Success);
-    if (result.status === CoreToolCallStatus.Success) {
+      expect(response).toEqual({
+        output: 'TruncatedContent...',
+        outputFile: '/tmp/truncated_output.txt',
+      });
       expect(result.response.outputFile).toBe('/tmp/truncated_output.txt');
     }
   });
@@ -597,6 +536,228 @@ describe('ToolExecutor', () => {
     expect(result.status).toBe(CoreToolCallStatus.Success);
   });
 
+  it('should truncate large output and move file when fullOutputFilePath is provided', async () => {
+    // 1. Setup Config for Truncation
+    vi.spyOn(config, 'getTruncateToolOutputThreshold').mockReturnValue(10);
+    vi.spyOn(config.storage, 'getProjectTempDir').mockReturnValue('/tmp');
+    vi.spyOn(fileUtils, 'moveToolOutputToFile').mockResolvedValue({
+      outputFile: '/tmp/moved_output.txt',
+    });
+
+    const mockTool = new MockTool({ name: SHELL_TOOL_NAME });
+    const invocation = mockTool.build({});
+    const longOutput = 'This is a very long output that should be truncated.';
+
+    // 2. Mock execution returning long content AND fullOutputFilePath
+    vi.mocked(coreToolHookTriggers.executeToolWithHooks).mockResolvedValue({
+      llmContent: longOutput,
+      returnDisplay: longOutput,
+      fullOutputFilePath: '/tmp/temp_full_output.txt',
+    });
+
+    const scheduledCall: ScheduledToolCall = {
+      status: CoreToolCallStatus.Scheduled,
+      request: {
+        callId: 'call-trunc-full',
+        name: SHELL_TOOL_NAME,
+        args: { command: 'echo long' },
+        isClientInitiated: false,
+        prompt_id: 'prompt-trunc-full',
+      },
+      tool: mockTool,
+      invocation: invocation as unknown as AnyToolInvocation,
+      startTime: Date.now(),
+    };
+
+    // 3. Execute
+    const result = await executor.execute({
+      call: scheduledCall,
+      signal: new AbortController().signal,
+      onUpdateToolCall: vi.fn(),
+    });
+
+    // 4. Verify Truncation Logic
+    expect(fileUtils.moveToolOutputToFile).toHaveBeenCalledWith(
+      '/tmp/temp_full_output.txt',
+      SHELL_TOOL_NAME,
+      'call-trunc-full',
+      expect.any(String), // temp dir
+      'test-session-id', // session id from makeFakeConfig
+    );
+
+    expect(fileUtils.formatTruncatedToolOutput).toHaveBeenCalledWith(
+      longOutput,
+      '/tmp/moved_output.txt',
+      10, // threshold (maxChars)
+    );
+
+    expect(result.status).toBe(CoreToolCallStatus.Success);
+    if (result.status === CoreToolCallStatus.Success) {
+      const response = result.response.responseParts[0]?.functionResponse
+        ?.response as Record<string, unknown>;
+      // The content should be the *truncated* version returned by the mock formatTruncatedToolOutput
+      expect(response).toEqual({
+        output: 'TruncatedContent...',
+        outputFile: '/tmp/moved_output.txt',
+      });
+      expect(result.response.outputFile).toBe('/tmp/moved_output.txt');
+    }
+  });
+
+  it('should preserve temporary file when fullOutputFilePath is provided but output is not truncated', async () => {
+    // 1. Setup Config for Truncation
+    vi.spyOn(config, 'getTruncateToolOutputThreshold').mockReturnValue(100);
+    vi.spyOn(config.storage, 'getProjectTempDir').mockReturnValue('/tmp');
+    vi.spyOn(config, 'getSessionId').mockReturnValue('test-session-id');
+    const unlinkSpy = vi
+      .spyOn(fsPromises, 'unlink')
+      .mockResolvedValue(undefined);
+    vi.spyOn(fileUtils, 'moveToolOutputToFile').mockResolvedValue({
+      outputFile: '/tmp/moved_output_short.txt',
+    });
+
+    const mockTool = new MockTool({ name: SHELL_TOOL_NAME });
+    const invocation = mockTool.build({});
+    const shortOutput = 'Short';
+
+    // 2. Mock execution returning short content AND fullOutputFilePath
+    vi.mocked(coreToolHookTriggers.executeToolWithHooks).mockResolvedValue({
+      llmContent: shortOutput,
+      returnDisplay: shortOutput,
+      fullOutputFilePath: '/tmp/temp_full_output_short.txt',
+    });
+
+    const scheduledCall: ScheduledToolCall = {
+      status: CoreToolCallStatus.Scheduled,
+      request: {
+        callId: 'call-short-full',
+        name: SHELL_TOOL_NAME,
+        args: { command: 'echo short' },
+        isClientInitiated: false,
+        prompt_id: 'prompt-short-full',
+      },
+      tool: mockTool,
+      invocation: invocation as unknown as AnyToolInvocation,
+      startTime: Date.now(),
+    };
+
+    // 3. Execute
+    const result = await executor.execute({
+      call: scheduledCall,
+      signal: new AbortController().signal,
+      onUpdateToolCall: vi.fn(),
+    });
+
+    // 4. Verify file preservation
+    expect(fileUtils.moveToolOutputToFile).toHaveBeenCalledWith(
+      '/tmp/temp_full_output_short.txt',
+      SHELL_TOOL_NAME,
+      'call-short-full',
+      expect.any(String),
+      'test-session-id',
+    );
+    expect(unlinkSpy).not.toHaveBeenCalled();
+    expect(fileUtils.formatTruncatedToolOutput).not.toHaveBeenCalled();
+
+    // We should save it since fullOutputFilePath was provided
+    expect(result.status).toBe(CoreToolCallStatus.Success);
+    if (result.status === CoreToolCallStatus.Success) {
+      const response = result.response.responseParts[0]?.functionResponse
+        ?.response as Record<string, unknown>;
+      expect(response).toEqual({
+        output: 'Short',
+        outputFile: '/tmp/moved_output_short.txt',
+      });
+      expect(result.response.outputFile).toBe('/tmp/moved_output_short.txt');
+    }
+
+    unlinkSpy.mockRestore();
+  });
+
+  it('should delete temporary file on error if fullOutputFilePath is provided', async () => {
+    const unlinkSpy = vi
+      .spyOn(fsPromises, 'unlink')
+      .mockResolvedValue(undefined);
+    const mockTool = new MockTool({ name: 'failTool' });
+    const invocation = mockTool.build({});
+
+    vi.mocked(coreToolHookTriggers.executeToolWithHooks).mockResolvedValue({
+      llmContent: 'partial',
+      returnDisplay: 'partial',
+      fullOutputFilePath: '/tmp/temp_error.txt',
+      error: { message: 'Tool Failed' },
+    });
+
+    const scheduledCall: ScheduledToolCall = {
+      status: CoreToolCallStatus.Scheduled,
+      request: {
+        callId: 'call-err',
+        name: 'failTool',
+        args: {},
+        isClientInitiated: false,
+        prompt_id: 'prompt-err',
+      },
+      tool: mockTool,
+      invocation: invocation as unknown as AnyToolInvocation,
+      startTime: Date.now(),
+    };
+
+    const result = await executor.execute({
+      call: scheduledCall,
+      signal: new AbortController().signal,
+      onUpdateToolCall: vi.fn(),
+    });
+
+    expect(unlinkSpy).toHaveBeenCalledWith('/tmp/temp_error.txt');
+    expect(result.status).toBe(CoreToolCallStatus.Error);
+    unlinkSpy.mockRestore();
+  });
+
+  it('should delete temporary file on abort if fullOutputFilePath is provided', async () => {
+    const unlinkSpy = vi
+      .spyOn(fsPromises, 'unlink')
+      .mockResolvedValue(undefined);
+    const mockTool = new MockTool({ name: 'slowTool' });
+    const invocation = mockTool.build({});
+
+    const controller = new AbortController();
+
+    vi.mocked(coreToolHookTriggers.executeToolWithHooks).mockImplementation(
+      async () => {
+        controller.abort();
+        return {
+          llmContent: 'partial',
+          returnDisplay: 'partial',
+          fullOutputFilePath: '/tmp/temp_abort.txt',
+        };
+      },
+    );
+
+    const scheduledCall: ScheduledToolCall = {
+      status: CoreToolCallStatus.Scheduled,
+      request: {
+        callId: 'call-abort',
+        name: 'slowTool',
+        args: {},
+        isClientInitiated: false,
+        prompt_id: 'prompt-abort',
+      },
+      tool: mockTool,
+      invocation: invocation as unknown as AnyToolInvocation,
+      startTime: Date.now(),
+    };
+
+    const result = await executor.execute({
+      call: scheduledCall,
+      signal: controller.signal,
+      onUpdateToolCall: vi.fn(),
+    });
+
+    expect(unlinkSpy).toHaveBeenCalledWith('/tmp/temp_abort.txt');
+    expect(result.status).toBe(CoreToolCallStatus.Cancelled);
+    unlinkSpy.mockRestore();
+  });
+
   it('should report execution ID updates for backgroundable tools', async () => {
     // 1. Setup ShellToolInvocation
     const messageBus = createMockMessageBus();

packages/core/src/scheduler/tool-executor.ts

@@ -4,6 +4,8 @@
  * SPDX-License-Identifier: Apache-2.0
  */
 
+import fsPromises from 'node:fs/promises';
+import { debugLogger } from '../utils/debugLogger.js';
 import {
   ToolErrorType,
   ToolOutputTruncatedEvent,
@@ -24,6 +26,7 @@ import { executeToolWithHooks } from '../core/coreToolHookTriggers.js';
 import {
   saveTruncatedToolOutput,
   formatTruncatedToolOutput,
+  moveToolOutputToFile,
 } from '../utils/fileUtils.js';
 import { convertToFunctionResponse } from '../utils/generateContentResponseUtilities.js';
 import {
@@ -138,6 +141,16 @@ export class ToolExecutor {
           }
 
           if (signal.aborted) {
+            if (toolResult.fullOutputFilePath) {
+              await fsPromises
+                .unlink(toolResult.fullOutputFilePath)
+                .catch((error) => {
+                  debugLogger.warn(
+                    `Failed to delete temporary tool output file on abort: ${toolResult.fullOutputFilePath}`,
+                    error,
+                  );
+                });
+            }
             completedToolCall = await this.createCancelledResult(
               call,
               'User cancelled tool execution.',
@@ -149,6 +162,16 @@ export class ToolExecutor {
               toolResult,
             );
           } else {
+            if (toolResult.fullOutputFilePath) {
+              await fsPromises
+                .unlink(toolResult.fullOutputFilePath)
+                .catch((error) => {
+                  debugLogger.warn(
+                    `Failed to delete temporary tool output file on error: ${toolResult.fullOutputFilePath}`,
+                    error,
+                  );
+                });
+            }
             const displayText =
               typeof toolResult.returnDisplay === 'string'
                 ? toolResult.returnDisplay
@@ -197,33 +220,80 @@ export class ToolExecutor {
   private async truncateOutputIfNeeded(
     call: ToolCall,
     content: PartListUnion,
+    fullOutputFilePath?: string,
   ): Promise<{ truncatedContent: PartListUnion; outputFile?: string }> {
+    const toolName = call.request.name;
+    const callId = call.request.callId;
+
     if (this.config.isContextManagementEnabled()) {
       const distiller = new ToolOutputDistillationService(
         this.config,
         this.context.geminiClient,
         this.context.promptId,
       );
-      return distiller.distill(call.request.name, call.request.callId, content);
+      const result = await distiller.distill(
+        call.request.name,
+        call.request.callId,
+        content,
+      );
+
+      let finalOutputFile = result.outputFile;
+      if (fullOutputFilePath) {
+        const { outputFile: movedPath } = await moveToolOutputToFile(
+          fullOutputFilePath,
+          toolName,
+          callId,
+          this.config.storage.getProjectTempDir(),
+          this.config.getSessionId(),
+        );
+        finalOutputFile = movedPath;
+        if (result.outputFile) {
+          try {
+            await fsPromises.unlink(result.outputFile);
+          } catch (error) {
+            debugLogger.warn(
+              `Failed to delete distiller's temporary tool output file: ${result.outputFile}`,
+              error,
+            );
+          }
+        }
+      }
+      return {
+        truncatedContent: result.truncatedContent,
+        outputFile: finalOutputFile,
+      };
     }
 
-    const toolName = call.request.name;
-    const callId = call.request.callId;
     let outputFile: string | undefined;
 
+    if (fullOutputFilePath) {
+      const { outputFile: movedPath } = await moveToolOutputToFile(
+        fullOutputFilePath,
+        toolName,
+        callId,
+        this.config.storage.getProjectTempDir(),
+        this.config.getSessionId(),
+      );
+      outputFile = movedPath;
+    }
+
     if (typeof content === 'string' && toolName === SHELL_TOOL_NAME) {
       const threshold = this.config.getTruncateToolOutputThreshold();
 
       if (threshold > 0 && content.length > threshold) {
         const originalContentLength = content.length;
-        const { outputFile: savedPath } = await saveTruncatedToolOutput(
-          content,
-          toolName,
-          callId,
-          this.config.storage.getProjectTempDir(),
-          this.context.promptId,
-        );
-        outputFile = savedPath;
+
+        if (!outputFile) {
+          const { outputFile: writtenPath } = await saveTruncatedToolOutput(
+            content,
+            toolName,
+            callId,
+            this.config.storage.getProjectTempDir(),
+            this.context.promptId,
+          );
+          outputFile = writtenPath;
+        }
+
         const truncatedContent = formatTruncatedToolOutput(
           content,
           outputFile,
@@ -255,14 +325,16 @@ export class ToolExecutor {
 
         if (threshold > 0 && textContent.length > threshold) {
           const originalContentLength = textContent.length;
-          const { outputFile: savedPath } = await saveTruncatedToolOutput(
-            textContent,
-            toolName,
-            callId,
-            this.config.storage.getProjectTempDir(),
-            this.context.promptId,
-          );
-          outputFile = savedPath;
+          if (!outputFile) {
+            const { outputFile: savedPath } = await saveTruncatedToolOutput(
+              textContent,
+              toolName,
+              callId,
+              this.config.storage.getProjectTempDir(),
+              this.context.promptId,
+            );
+            outputFile = savedPath;
+          }
           const truncatedText = formatTruncatedToolOutput(
             textContent,
             outputFile,
@@ -314,7 +386,11 @@ export class ToolExecutor {
       // Attempt to truncate and save output if we have content, even in cancellation case
       // This is to handle cases where the tool may have produced output before cancellation
       const { truncatedContent: output, outputFile: truncatedOutputFile } =
-        await this.truncateOutputIfNeeded(call, toolResult?.llmContent);
+        await this.truncateOutputIfNeeded(
+          call,
+          toolResult.llmContent,
+          toolResult.fullOutputFilePath,
+        );
 
       outputFile = truncatedOutputFile;
       responseParts = convertToFunctionResponse(
@@ -323,6 +399,7 @@ export class ToolExecutor {
         output,
         this.config.getActiveModel(),
         this.config,
+        outputFile,
       );
 
       // Inject the cancellation error into the response object
@@ -332,6 +409,16 @@ export class ToolExecutor {
         respObj['error'] = errorMessage;
       }
     } else {
+      if (toolResult?.fullOutputFilePath) {
+        try {
+          await fsPromises.unlink(toolResult.fullOutputFilePath);
+        } catch (error) {
+          debugLogger.warn(
+            `Failed to delete temporary tool output file: ${toolResult.fullOutputFilePath}`,
+            error,
+          );
+        }
+      }
       responseParts = [
         {
           functionResponse: {
@@ -369,8 +456,11 @@ export class ToolExecutor {
     toolResult: ToolResult,
   ): Promise<SuccessfulToolCall> {
     const { truncatedContent: content, outputFile } =
-      await this.truncateOutputIfNeeded(call, toolResult.llmContent);
-
+      await this.truncateOutputIfNeeded(
+        call,
+        toolResult.llmContent,
+        toolResult.fullOutputFilePath,
+      );
     const toolName = call.request.originalRequestName || call.request.name;
     const callId = call.request.callId;
 
@@ -380,6 +470,7 @@ export class ToolExecutor {
       content,
       this.config.getActiveModel(),
       this.config,
+      outputFile,
     );
 
     const successResponse: ToolCallResponseInfo = {

packages/core/src/services/executionLifecycleService.ts

@@ -27,6 +27,7 @@ export interface ExecutionResult {
   pid: number | undefined;
   executionMethod: ExecutionMethod;
   backgrounded?: boolean;
+  fullOutputFilePath?: string;
 }
 
 export interface ExecutionHandle {
@@ -35,6 +36,14 @@ export interface ExecutionHandle {
 }
 
 export type ExecutionOutputEvent =
+  | {
+      type: 'raw_data';
+      chunk: string;
+    }
+  | {
+      type: 'file_data';
+      chunk: string;
+    }
   | {
       type: 'data';
       chunk: string | AnsiOutput;

packages/core/src/services/keychainService.test.ts

@@ -53,7 +53,7 @@ vi.mock('../utils/events.js', () => ({
 }));
 
 vi.mock('../utils/debugLogger.js', () => ({
-  debugLogger: { log: vi.fn() },
+  debugLogger: { debug: vi.fn() },
 }));
 
 vi.mock('node:os', async (importOriginal) => {
@@ -153,14 +153,14 @@ describe('KeychainService', () => {
 
       // Because it falls back to FileKeychain, it is always available.
       expect(available).toBe(true);
-      expect(debugLogger.log).toHaveBeenCalledWith(
+      expect(debugLogger.debug).toHaveBeenCalledWith(
         expect.stringContaining('encountered an error'),
         'locked',
       );
       expect(coreEvents.emitTelemetryKeychainAvailability).toHaveBeenCalledWith(
         expect.objectContaining({ available: false }),
       );
-      expect(debugLogger.log).toHaveBeenCalledWith(
+      expect(debugLogger.debug).toHaveBeenCalledWith(
         expect.stringContaining('Using FileKeychain fallback'),
       );
       expect(FileKeychain).toHaveBeenCalled();
@@ -173,7 +173,7 @@ describe('KeychainService', () => {
       const available = await service.isAvailable();
 
       expect(available).toBe(true);
-      expect(debugLogger.log).toHaveBeenCalledWith(
+      expect(debugLogger.debug).toHaveBeenCalledWith(
         expect.stringContaining('failed structural validation'),
         expect.objectContaining({ getPassword: expect.any(Array) }),
       );
@@ -191,7 +191,7 @@ describe('KeychainService', () => {
       const available = await service.isAvailable();
 
       expect(available).toBe(true);
-      expect(debugLogger.log).toHaveBeenCalledWith(
+      expect(debugLogger.debug).toHaveBeenCalledWith(
         expect.stringContaining('functional verification failed'),
       );
       expect(FileKeychain).toHaveBeenCalled();
@@ -243,7 +243,7 @@ describe('KeychainService', () => {
       );
       expect(mockKeytar.setPassword).not.toHaveBeenCalled();
       expect(FileKeychain).toHaveBeenCalled();
-      expect(debugLogger.log).toHaveBeenCalledWith(
+      expect(debugLogger.debug).toHaveBeenCalledWith(
         expect.stringContaining('MacOS default keychain not found'),
       );
     });

packages/core/src/services/keychainService.ts

@@ -114,7 +114,7 @@ export class KeychainService {
     }
 
     // If native failed or was skipped, return the secure file fallback.
-    debugLogger.log('Using FileKeychain fallback for secure storage.');
+    debugLogger.debug('Using FileKeychain fallback for secure storage.');
     return new FileKeychain();
   }
 
@@ -130,7 +130,7 @@ export class KeychainService {
 
       // Probing macOS prevents process-blocking popups when no keychain exists.
       if (os.platform() === 'darwin' && !this.isMacOSKeychainAvailable()) {
-        debugLogger.log(
+        debugLogger.debug(
           'MacOS default keychain not found; skipping functional verification.',
         );
         return null;
@@ -140,12 +140,15 @@ export class KeychainService {
         return keychainModule;
       }
 
-      debugLogger.log('Keychain functional verification failed');
+      debugLogger.debug('Keychain functional verification failed');
       return null;
     } catch (error) {
       // Avoid logging full error objects to prevent PII exposure.
       const message = error instanceof Error ? error.message : String(error);
-      debugLogger.log('Keychain initialization encountered an error:', message);
+      debugLogger.debug(
+        'Keychain initialization encountered an error:',
+        message,
+      );
       return null;
     }
   }
@@ -162,7 +165,7 @@ export class KeychainService {
       return potential as Keychain;
     }
 
-    debugLogger.log(
+    debugLogger.debug(
       'Keychain module failed structural validation:',
       result.error.flatten().fieldErrors,
     );

packages/core/src/services/shellExecutionService.test.ts

@@ -33,7 +33,16 @@ const mockIsBinary = vi.hoisted(() => vi.fn());
 const mockPlatform = vi.hoisted(() => vi.fn());
 const mockHomedir = vi.hoisted(() => vi.fn());
 const mockMkdirSync = vi.hoisted(() => vi.fn());
-const mockCreateWriteStream = vi.hoisted(() => vi.fn());
+const mockCreateWriteStream = vi.hoisted(() =>
+  vi.fn().mockReturnValue({
+    write: vi.fn(),
+    end: vi.fn().mockImplementation((cb?: () => void) => {
+      if (cb) cb();
+    }),
+    destroy: vi.fn(),
+    closed: false,
+  }),
+);
 const mockGetPty = vi.hoisted(() => vi.fn());
 const mockSerializeTerminalToObject = vi.hoisted(() => vi.fn());
 const mockResolveExecutable = vi.hoisted(() => vi.fn());
@@ -92,6 +101,7 @@ vi.mock('node:os', () => ({
   default: {
     platform: mockPlatform,
     homedir: mockHomedir,
+    tmpdir: () => '/tmp',
     constants: {
       signals: {
         SIGTERM: 15,
@@ -208,6 +218,15 @@ describe('ShellExecutionService', () => {
 
   beforeEach(() => {
     vi.clearAllMocks();
+    mockCreateWriteStream.mockReturnValue({
+      write: vi.fn(),
+      end: vi.fn().mockImplementation((cb?: () => void) => {
+        if (cb) cb();
+      }),
+      destroy: vi.fn(),
+      closed: false,
+      on: vi.fn(),
+    });
     ExecutionLifecycleService.resetForTest();
     ShellExecutionService.resetForTest();
     mockSerializeTerminalToObject.mockReturnValue([]);
@@ -489,6 +508,7 @@ describe('ShellExecutionService', () => {
           // eslint-disable-next-line @typescript-eslint/no-explicit-any
           headlessTerminal: mockHeadlessTerminal as any,
           command: 'some-command',
+          lastCommittedLine: -1,
         });
     });
 
@@ -620,7 +640,7 @@ describe('ShellExecutionService', () => {
     });
 
     it('should handle a synchronous spawn error', async () => {
-      mockGetPty.mockImplementation(() => null);
+      mockGetPty.mockResolvedValue(null);
 
       mockCpSpawn.mockImplementation(() => {
         throw new Error('Simulated PTY spawn error');
@@ -724,7 +744,13 @@ describe('ShellExecutionService', () => {
   });
 
   describe('Backgrounding', () => {
-    let mockWriteStream: { write: Mock; end: Mock; on: Mock };
+    let mockWriteStream: {
+      write: Mock;
+      end: Mock;
+      on: Mock;
+      destroy: Mock;
+      closed: boolean;
+    };
     let mockBgChildProcess: EventEmitter & Partial<ChildProcess>;
 
     beforeEach(async () => {
@@ -732,6 +758,8 @@ describe('ShellExecutionService', () => {
         write: vi.fn(),
         end: vi.fn().mockImplementation((cb) => cb?.()),
         on: vi.fn(),
+        destroy: vi.fn(),
+        closed: false,
       };
 
       mockMkdirSync.mockReturnValue(undefined);
@@ -989,6 +1017,8 @@ describe('ShellExecutionService', () => {
         pty.onExit.mock.calls[0][0]({ exitCode: 0, signal: null });
       });
 
+      // We don't check result here because result is not available in the test body as written
+      // The test body doesn't capture the return value of simulateExecution correctly for this assertion.
       expect(onOutputEventMock).toHaveBeenCalledTimes(4);
       expect(onOutputEventMock.mock.calls[0][0]).toEqual({
         type: 'binary_detected',
@@ -1301,7 +1331,7 @@ describe('ShellExecutionService child_process fallback', () => {
       const { result, handle } = await simulateExecution('ls -l', (cp) => {
         cp.stdout?.emit('data', Buffer.from('file1.txt\n'));
         cp.stderr?.emit('data', Buffer.from('a warning'));
-        cp.emit('exit', 0, null);
+        cp.emit('close', 0, null);
         cp.emit('close', 0, null);
       });
 
@@ -1338,7 +1368,7 @@ describe('ShellExecutionService child_process fallback', () => {
     it('should strip ANSI color codes from output', async () => {
       const { result } = await simulateExecution('ls --color=auto', (cp) => {
         cp.stdout?.emit('data', Buffer.from('a\u001b[31mred\u001b[0mword'));
-        cp.emit('exit', 0, null);
+        cp.emit('close', 0, null);
         cp.emit('close', 0, null);
       });
 
@@ -1359,7 +1389,7 @@ describe('ShellExecutionService child_process fallback', () => {
         const multiByteChar = Buffer.from('你好', 'utf-8');
         cp.stdout?.emit('data', multiByteChar.slice(0, 2));
         cp.stdout?.emit('data', multiByteChar.slice(2));
-        cp.emit('exit', 0, null);
+        cp.emit('close', 0, null);
         cp.emit('close', 0, null);
       });
       expect(result.output.trim()).toBe('你好');
@@ -1367,7 +1397,7 @@ describe('ShellExecutionService child_process fallback', () => {
 
     it('should handle commands with no output', async () => {
       const { result } = await simulateExecution('touch file', (cp) => {
-        cp.emit('exit', 0, null);
+        cp.emit('close', 0, null);
         cp.emit('close', 0, null);
       });
 
@@ -1389,7 +1419,7 @@ describe('ShellExecutionService child_process fallback', () => {
         cp.stdout?.emit('data', Buffer.from(chunk1));
         cp.stdout?.emit('data', Buffer.from(chunk2));
         cp.stdout?.emit('data', Buffer.from(chunk3));
-        cp.emit('exit', 0, null);
+        cp.emit('close', 0, null);
       });
 
       const truncationMessage =
@@ -1414,7 +1444,7 @@ describe('ShellExecutionService child_process fallback', () => {
     it('should capture a non-zero exit code and format output correctly', async () => {
       const { result } = await simulateExecution('a-bad-command', (cp) => {
         cp.stderr?.emit('data', Buffer.from('command not found'));
-        cp.emit('exit', 127, null);
+        cp.emit('close', 127, null);
         cp.emit('close', 127, null);
       });
 
@@ -1425,7 +1455,7 @@ describe('ShellExecutionService child_process fallback', () => {
 
     it('should capture a termination signal', async () => {
       const { result } = await simulateExecution('long-process', (cp) => {
-        cp.emit('exit', null, 'SIGTERM');
+        cp.emit('close', null, 'SIGTERM');
         cp.emit('close', null, 'SIGTERM');
       });
 
@@ -1437,7 +1467,7 @@ describe('ShellExecutionService child_process fallback', () => {
       const spawnError = new Error('spawn EACCES');
       const { result } = await simulateExecution('protected-cmd', (cp) => {
         cp.emit('error', spawnError);
-        cp.emit('exit', 1, null);
+        cp.emit('close', 1, null);
         cp.emit('close', 1, null);
       });
 
@@ -1483,11 +1513,11 @@ describe('ShellExecutionService child_process fallback', () => {
               await new Promise(process.nextTick);
               await new Promise(process.nextTick);
               if (expectedExit.signal) {
-                cp.emit('exit', null, expectedExit.signal);
+                cp.emit('close', null, expectedExit.signal);
                 cp.emit('close', null, expectedExit.signal);
               }
               if (typeof expectedExit.code === 'number') {
-                cp.emit('exit', expectedExit.code, null);
+                cp.emit('close', expectedExit.code, null);
                 cp.emit('close', expectedExit.code, null);
               }
             },
@@ -1556,7 +1586,7 @@ describe('ShellExecutionService child_process fallback', () => {
       );
 
       // Finally, simulate the process exiting and await the result
-      mockChildProcess.emit('exit', null, 'SIGKILL');
+      mockChildProcess.emit('close', null, 'SIGKILL');
       mockChildProcess.emit('close', null, 'SIGKILL');
       const result = await handle.result;
 
@@ -1576,9 +1606,11 @@ describe('ShellExecutionService child_process fallback', () => {
       await simulateExecution('cat image.png', (cp) => {
         cp.stdout?.emit('data', binaryChunk1);
         cp.stdout?.emit('data', binaryChunk2);
-        cp.emit('exit', 0, null);
+        cp.emit('close', 0, null);
       });
 
+      // We don't check result here because result is not available in the test body as written
+      // The test body doesn't capture the return value of simulateExecution correctly for this assertion.
       expect(onOutputEventMock).toHaveBeenCalledTimes(4);
       expect(onOutputEventMock.mock.calls[0][0]).toEqual({
         type: 'binary_detected',
@@ -1604,7 +1636,6 @@ describe('ShellExecutionService child_process fallback', () => {
       await simulateExecution('cat mixed_file', (cp) => {
         cp.stdout?.emit('data', Buffer.from([0x00, 0x01, 0x02]));
         cp.stdout?.emit('data', Buffer.from('more text'));
-        cp.emit('exit', 0, null);
         cp.emit('close', 0, null);
       });
 
@@ -1640,7 +1671,7 @@ describe('ShellExecutionService child_process fallback', () => {
     it('should use powershell.exe on Windows', async () => {
       mockPlatform.mockReturnValue('win32');
       await simulateExecution('dir "foo bar"', (cp) => {
-        cp.emit('exit', 0, null);
+        cp.emit('close', 0, null);
       });
 
       expect(mockCpSpawn).toHaveBeenCalledWith(
@@ -1657,7 +1688,7 @@ describe('ShellExecutionService child_process fallback', () => {
     it('should use bash and detached process group on Linux', async () => {
       mockPlatform.mockReturnValue('linux');
       await simulateExecution('ls "foo bar"', (cp) => {
-        cp.emit('exit', 0, null);
+        cp.emit('close', 0, null);
       });
 
       expect(mockCpSpawn).toHaveBeenCalledWith(
@@ -1771,7 +1802,7 @@ describe('ShellExecutionService execution method selection', () => {
     );
 
     // Simulate exit to allow promise to resolve
-    mockChildProcess.emit('exit', 0, null);
+    mockChildProcess.emit('close', 0, null);
     const result = await handle.result;
 
     expect(mockGetPty).not.toHaveBeenCalled();
@@ -1794,7 +1825,7 @@ describe('ShellExecutionService execution method selection', () => {
     );
 
     // Simulate exit to allow promise to resolve
-    mockChildProcess.emit('exit', 0, null);
+    mockChildProcess.emit('close', 0, null);
     const result = await handle.result;
 
     expect(mockGetPty).toHaveBeenCalled();
@@ -1860,10 +1891,6 @@ describe('ShellExecutionService environment variables', () => {
       // Small delay to allow async ops to complete
       setTimeout(() => mockPtyProcess.emit('exit', { exitCode, signal }), 0);
     });
-    mockChildProcess.on('exit', (code, signal) => {
-      // Small delay to allow async ops to complete
-      setTimeout(() => mockChildProcess.emit('close', code, signal), 0);
-    });
   });
 
   afterEach(() => {
@@ -1926,7 +1953,7 @@ describe('ShellExecutionService environment variables', () => {
     expect(cpEnv).toHaveProperty('GEMINI_CLI_TEST_VAR', 'test-value');
 
     // Ensure child_process exits
-    mockChildProcess.emit('exit', 0, null);
+    mockChildProcess.emit('close', 0, null);
     mockChildProcess.emit('close', 0, null);
     await new Promise(process.nextTick);
   });
@@ -1986,7 +2013,7 @@ describe('ShellExecutionService environment variables', () => {
     expect(cpEnv).toHaveProperty('GEMINI_CLI_TEST_VAR', 'test-value');
 
     // Ensure child_process exits
-    mockChildProcess.emit('exit', 0, null);
+    mockChildProcess.emit('close', 0, null);
     mockChildProcess.emit('close', 0, null);
     await new Promise(process.nextTick);
   });
@@ -2034,7 +2061,7 @@ describe('ShellExecutionService environment variables', () => {
     expect(cpEnv).toHaveProperty('GEMINI_CLI', '1');
 
     // Ensure child_process exits
-    mockChildProcess.emit('exit', 0, null);
+    mockChildProcess.emit('close', 0, null);
     mockChildProcess.emit('close', 0, null);
     await new Promise(process.nextTick);
   });
@@ -2091,7 +2118,7 @@ describe('ShellExecutionService environment variables', () => {
     );
 
     // Clean up
-    mockChild.emit('exit', 0, null);
+    mockChild.emit('close', 0, null);
     mockChild.emit('close', 0, null);
     await handle.result;
   });
@@ -2140,7 +2167,7 @@ describe('ShellExecutionService environment variables', () => {
     expect(cpEnv).toHaveProperty('GIT_CONFIG_VALUE_2', '');
 
     // Ensure child_process exits
-    mockChildProcess.emit('exit', 0, null);
+    mockChildProcess.emit('close', 0, null);
     mockChildProcess.emit('close', 0, null);
     await new Promise(process.nextTick);
 
@@ -2180,7 +2207,7 @@ describe('ShellExecutionService environment variables', () => {
     expect(cpEnv).not.toHaveProperty('GIT_CONFIG_COUNT');
 
     // Ensure child_process exits
-    mockChildProcess.emit('exit', 0, null);
+    mockChildProcess.emit('close', 0, null);
     mockChildProcess.emit('close', 0, null);
     await new Promise(process.nextTick);
 

packages/core/src/services/shellExecutionService.ts

@@ -12,6 +12,7 @@ import type { Writable } from 'node:stream';
 import os from 'node:os';
 import fs, { mkdirSync } from 'node:fs';
 import path from 'node:path';
+import crypto from 'node:crypto';
 import type { IPty } from '@lydell/node-pty';
 import { getCachedEncodingForBuffer } from '../utils/systemEncoding.js';
 import {
@@ -120,6 +121,8 @@ interface ActivePty {
   maxSerializedLines?: number;
   command: string;
   sessionId?: string;
+  lastSerializedOutput?: AnsiOutput;
+  lastCommittedLine: number;
 }
 
 interface ActiveChildProcess {
@@ -148,6 +151,48 @@ const findLastContentLine = (
   return -1;
 };
 
+const emitPendingLines = (
+  activePty: ActivePty,
+  pid: number,
+  onOutputEvent: (event: ShellOutputEvent) => void,
+  forceAll = false,
+) => {
+  const buffer = activePty.headlessTerminal.buffer.active;
+  const limit = forceAll ? buffer.length : buffer.baseY;
+
+  let chunks = '';
+  for (let i = activePty.lastCommittedLine + 1; i < limit; i++) {
+    const line = buffer.getLine(i);
+    if (!line) continue;
+
+    let trimRight = true;
+    let isNextLineWrapped = false;
+    if (i + 1 < buffer.length) {
+      const nextLine = buffer.getLine(i + 1);
+      if (nextLine?.isWrapped) {
+        isNextLineWrapped = true;
+        trimRight = false;
+      }
+    }
+
+    const lineContent = line.translateToString(trimRight);
+    chunks += lineContent;
+    if (!isNextLineWrapped) {
+      chunks += '\n';
+    }
+  }
+
+  if (chunks.length > 0) {
+    const event: ShellOutputEvent = {
+      type: 'file_data',
+      chunk: chunks,
+    };
+    onOutputEvent(event);
+    ExecutionLifecycleService.emitEvent(pid, event);
+    activePty.lastCommittedLine = limit - 1;
+  }
+};
+
 const getFullBufferText = (terminal: pkg.Terminal, startLine = 0): string => {
   const buffer = terminal.buffer.active;
   const lines: string[] = [];
@@ -326,32 +371,110 @@ export class ShellExecutionService {
     shouldUseNodePty: boolean,
     shellExecutionConfig: ShellExecutionConfig,
   ): Promise<ShellExecutionHandle> {
+    const outputFileName = `gemini_shell_output_${crypto.randomBytes(6).toString('hex')}.log`;
+    const outputFilePath = path.join(os.tmpdir(), outputFileName);
+    const outputStream = fs.createWriteStream(outputFilePath);
+
+    let totalBytesWritten = 0;
+
+    const interceptedOnOutputEvent = (event: ShellOutputEvent) => {
+      switch (event.type) {
+        case 'raw_data':
+          break;
+        case 'file_data':
+          outputStream.write(event.chunk);
+          totalBytesWritten += Buffer.byteLength(event.chunk);
+          break;
+        case 'binary_detected':
+        case 'binary_progress':
+          break;
+        default:
+          break;
+      }
+      onOutputEvent(event);
+    };
+
+    let handlePromise: Promise<ShellExecutionHandle>;
+
     if (shouldUseNodePty) {
-      const ptyInfo = await getPty();
-      if (ptyInfo) {
-        try {
-          return await this.executeWithPty(
+      handlePromise = getPty().then((ptyInfo) => {
+        if (ptyInfo) {
+          return this.executeWithPty(
             commandToExecute,
             cwd,
-            onOutputEvent,
+            interceptedOnOutputEvent,
             abortSignal,
             shellExecutionConfig,
             ptyInfo,
+          ).catch(() =>
+            this.childProcessFallback(
+              commandToExecute,
+              cwd,
+              interceptedOnOutputEvent,
+              abortSignal,
+              shellExecutionConfig,
+              shouldUseNodePty,
+            ),
           );
-        } catch {
-          // Fallback to child_process
         }
-      }
+        return this.childProcessFallback(
+          commandToExecute,
+          cwd,
+          interceptedOnOutputEvent,
+          abortSignal,
+          shellExecutionConfig,
+          shouldUseNodePty,
+        );
+      });
+    } else {
+      handlePromise = this.childProcessFallback(
+        commandToExecute,
+        cwd,
+        interceptedOnOutputEvent,
+        abortSignal,
+        shellExecutionConfig,
+        shouldUseNodePty,
+      );
     }
 
-    return this.childProcessFallback(
-      commandToExecute,
-      cwd,
-      onOutputEvent,
-      abortSignal,
-      shellExecutionConfig,
-      shouldUseNodePty,
-    );
+    const handle = await handlePromise;
+
+    const wrappedResultPromise = handle.result
+      .then(async (result) => {
+        await new Promise<void>((resolve) => {
+          outputStream.end(resolve);
+        });
+        // The threshold logic is handled later by ToolExecutor/caller, so we just return the full file path if anything was written
+        if (
+          totalBytesWritten > 0 &&
+          !result.backgrounded &&
+          !abortSignal.aborted &&
+          !result.error
+        ) {
+          return {
+            ...result,
+            fullOutputFilePath: outputFilePath,
+          };
+        } else {
+          if (!outputStream.closed) {
+            outputStream.destroy();
+          }
+          await fs.promises.unlink(outputFilePath).catch(() => undefined);
+          return result;
+        }
+      })
+      .catch(async (err) => {
+        if (!outputStream.closed) {
+          outputStream.destroy();
+        }
+        await fs.promises.unlink(outputFilePath).catch(() => undefined);
+        throw err;
+      });
+
+    return {
+      pid: handle.pid,
+      result: wrappedResultPromise,
+    };
   }
 
   private static appendAndTruncate(
@@ -661,6 +784,24 @@ export class ShellExecutionService {
           }
 
           if (decodedChunk) {
+            const rawEvent: ShellOutputEvent = {
+              type: 'raw_data',
+              chunk: decodedChunk,
+            };
+            onOutputEvent(rawEvent);
+            if (child.pid) {
+              ExecutionLifecycleService.emitEvent(child.pid, rawEvent);
+            }
+
+            const fileEvent: ShellOutputEvent = {
+              type: 'file_data',
+              chunk: stripAnsi(decodedChunk),
+            };
+            onOutputEvent(fileEvent);
+            if (child.pid) {
+              ExecutionLifecycleService.emitEvent(child.pid, fileEvent);
+            }
+
             const event: ShellOutputEvent = {
               type: 'data',
               chunk: decodedChunk,
@@ -772,7 +913,7 @@ export class ShellExecutionService {
 
       abortSignal.addEventListener('abort', abortHandler, { once: true });
 
-      child.on('exit', (code, signal) => {
+      child.on('close', (code, signal) => {
         handleExit(code, signal);
       });
 
@@ -784,6 +925,15 @@ export class ShellExecutionService {
           if (remaining) {
             state.output += remaining;
             if (isStreamingRawContent) {
+              const rawEvent: ShellOutputEvent = {
+                type: 'raw_data',
+                chunk: remaining,
+              };
+              onOutputEvent(rawEvent);
+              if (child.pid) {
+                ExecutionLifecycleService.emitEvent(child.pid, rawEvent);
+              }
+
               const event: ShellOutputEvent = {
                 type: 'data',
                 chunk: remaining,
@@ -792,6 +942,15 @@ export class ShellExecutionService {
               if (child.pid) {
                 ExecutionLifecycleService.emitEvent(child.pid, event);
               }
+
+              const fileEvent: ShellOutputEvent = {
+                type: 'file_data',
+                chunk: stripAnsi(remaining),
+              };
+              onOutputEvent(fileEvent);
+              if (child.pid) {
+                ExecutionLifecycleService.emitEvent(child.pid, fileEvent);
+              }
             }
           }
         }
@@ -800,6 +959,15 @@ export class ShellExecutionService {
           if (remaining) {
             state.output += remaining;
             if (isStreamingRawContent) {
+              const rawEvent: ShellOutputEvent = {
+                type: 'raw_data',
+                chunk: remaining,
+              };
+              onOutputEvent(rawEvent);
+              if (child.pid) {
+                ExecutionLifecycleService.emitEvent(child.pid, rawEvent);
+              }
+
               const event: ShellOutputEvent = {
                 type: 'data',
                 chunk: remaining,
@@ -808,6 +976,15 @@ export class ShellExecutionService {
               if (child.pid) {
                 ExecutionLifecycleService.emitEvent(child.pid, event);
               }
+
+              const fileEvent: ShellOutputEvent = {
+                type: 'file_data',
+                chunk: stripAnsi(remaining),
+              };
+              onOutputEvent(fileEvent);
+              if (child.pid) {
+                ExecutionLifecycleService.emitEvent(child.pid, fileEvent);
+              }
             }
           }
         }
@@ -934,6 +1111,7 @@ export class ShellExecutionService {
         maxSerializedLines: shellExecutionConfig.maxSerializedLines,
         command: shellExecutionConfig.originalCommand ?? commandToExecute,
         sessionId: shellExecutionConfig.sessionId,
+        lastCommittedLine: -1,
       });
 
       const result = ExecutionLifecycleService.attachExecution(ptyPid, {
@@ -1099,10 +1277,38 @@ export class ShellExecutionService {
         }, 68);
       };
 
-      headlessTerminal.onScroll(() => {
+      let lastYdisp = 0;
+      let hasReachedMax = false;
+      const scrollbackLimit =
+        shellExecutionConfig.scrollback ?? SCROLLBACK_LIMIT;
+
+      headlessTerminal.onScroll((ydisp) => {
         if (!isWriting) {
           render();
         }
+
+        if (
+          ydisp === scrollbackLimit &&
+          lastYdisp === scrollbackLimit &&
+          hasReachedMax
+        ) {
+          const activePty = this.activePtys.get(ptyPid);
+          if (activePty) {
+            activePty.lastCommittedLine--;
+          }
+        }
+        if (
+          ydisp === scrollbackLimit &&
+          headlessTerminal.buffer.active.length === scrollbackLimit + rows
+        ) {
+          hasReachedMax = true;
+        }
+        lastYdisp = ydisp;
+
+        const activePtyForEmit = this.activePtys.get(ptyPid);
+        if (activePtyForEmit) {
+          emitPendingLines(activePtyForEmit, ptyPid, onOutputEvent);
+        }
       });
 
       const handleOutput = (data: Buffer) => {
@@ -1187,6 +1393,11 @@ export class ShellExecutionService {
             render(true);
             cmdCleanup?.();
 
+            const activePty = ShellExecutionService.activePtys.get(ptyPid);
+            if (activePty && isStreamingRawContent) {
+              emitPendingLines(activePty, ptyPid, onOutputEvent, true);
+            }
+
             const event: ShellOutputEvent = {
               type: 'exit',
               exitCode,
@@ -1490,6 +1701,7 @@ export class ShellExecutionService {
         startLine,
         endLine,
       );
+      activePty.lastSerializedOutput = bufferData;
       const event: ShellOutputEvent = { type: 'data', chunk: bufferData };
       ExecutionLifecycleService.emitEvent(pid, event);
     }

packages/core/src/telemetry/memory-monitor.test.ts

@@ -17,13 +17,18 @@ import {
   _resetGlobalMemoryMonitorForTests,
 } from './memory-monitor.js';
 import type { Config } from '../config/config.js';
-import { recordMemoryUsage, isPerformanceMonitoringActive } from './metrics.js';
+import {
+  recordMemoryUsage,
+  recordCpuUsage,
+  isPerformanceMonitoringActive,
+} from './metrics.js';
 import { HighWaterMarkTracker } from './high-water-mark-tracker.js';
 import { RateLimiter } from './rate-limiter.js';
 
 // Mock dependencies
 vi.mock('./metrics.js', () => ({
   recordMemoryUsage: vi.fn(),
+  recordCpuUsage: vi.fn(),
   isPerformanceMonitoringActive: vi.fn(),
   MemoryMetricType: {
     HEAP_USED: 'heap_used',
@@ -50,6 +55,7 @@ vi.mock('node:process', () => ({
 }));
 
 const mockRecordMemoryUsage = vi.mocked(recordMemoryUsage);
+const mockRecordCpuUsage = vi.mocked(recordCpuUsage);
 const mockIsPerformanceMonitoringActive = vi.mocked(
   isPerformanceMonitoringActive,
 );
@@ -192,6 +198,13 @@ describe('MemoryMonitor', () => {
             component: 'test_context',
           },
         );
+        expect(mockRecordCpuUsage).toHaveBeenCalledWith(
+          mockConfig,
+          expect.any(Number),
+          {
+            component: 'test_context',
+          },
+        );
       });
 
       it('should not record metrics when performance monitoring is inactive', () => {

packages/core/src/telemetry/memory-monitor.ts

@@ -12,6 +12,7 @@ import { isUserActive } from './activity-detector.js';
 import { HighWaterMarkTracker } from './high-water-mark-tracker.js';
 import {
   recordMemoryUsage,
+  recordCpuUsage,
   MemoryMetricType,
   isPerformanceMonitoringActive,
 } from './metrics.js';
@@ -37,6 +38,7 @@ export class MemoryMonitor {
   private intervalId: NodeJS.Timeout | null = null;
   private isRunning = false;
   private lastSnapshot: MemorySnapshot | null = null;
+  private lastCpuUsage: NodeJS.CpuUsage | null = null;
   private monitoringInterval: number = 10000;
   private highWaterMarkTracker: HighWaterMarkTracker;
   private rateLimiter: RateLimiter;
@@ -191,6 +193,13 @@ export class MemoryMonitor {
         memory_type: MemoryMetricType.RSS,
         component: context,
       });
+
+      // Record delta CPU usage (in microseconds)
+      const cpuUsage = process.cpuUsage(this.lastCpuUsage ?? undefined);
+      this.lastCpuUsage = process.cpuUsage();
+      recordCpuUsage(config, cpuUsage.user + cpuUsage.system, {
+        component: context,
+      });
     }
 
     this.lastSnapshot = snapshot;

packages/core/src/tools/ripGrep.test.ts

@@ -4,20 +4,13 @@
  * SPDX-License-Identifier: Apache-2.0
  */
 
-import {
-  describe,
-  it,
-  expect,
-  beforeEach,
-  afterEach,
-  afterAll,
-  vi,
-} from 'vitest';
+import { describe, it, expect, beforeEach, afterEach, vi } from 'vitest';
 import {
   canUseRipgrep,
   RipGrepTool,
   ensureRgPath,
   type RipGrepToolParams,
+  getRipgrepPath,
 } from './ripGrep.js';
 import type { GrepResult } from './tools.js';
 import path from 'node:path';
@@ -25,18 +18,21 @@ import { isSubpath } from '../utils/paths.js';
 import fs from 'node:fs/promises';
 import os from 'node:os';
 import type { Config } from '../config/config.js';
-import { Storage } from '../config/storage.js';
 import { GEMINI_IGNORE_FILE_NAME } from '../config/constants.js';
 import { createMockWorkspaceContext } from '../test-utils/mockWorkspaceContext.js';
 import { spawn, type ChildProcess } from 'node:child_process';
 import { PassThrough, Readable } from 'node:stream';
 import EventEmitter from 'node:events';
-import { downloadRipGrep } from '@joshua.litt/get-ripgrep';
 import { createMockMessageBus } from '../test-utils/mock-message-bus.js';
-// Mock dependencies for canUseRipgrep
-vi.mock('@joshua.litt/get-ripgrep', () => ({
-  downloadRipGrep: vi.fn(),
-}));
+import { fileExists } from '../utils/fileUtils.js';
+
+vi.mock('../utils/fileUtils.js', async (importOriginal) => {
+  const actual = await importOriginal<typeof import('../utils/fileUtils.js')>();
+  return {
+    ...actual,
+    fileExists: vi.fn(),
+  };
+});
 
 // Mock child_process for ripgrep calls
 vi.mock('child_process', () => ({
@@ -44,161 +40,42 @@ vi.mock('child_process', () => ({
 }));
 
 const mockSpawn = vi.mocked(spawn);
-const downloadRipGrepMock = vi.mocked(downloadRipGrep);
-const originalGetGlobalBinDir = Storage.getGlobalBinDir.bind(Storage);
-const storageSpy = vi.spyOn(Storage, 'getGlobalBinDir');
-
-function getRipgrepBinaryName() {
-  return process.platform === 'win32' ? 'rg.exe' : 'rg';
-}
 
 describe('canUseRipgrep', () => {
-  let tempRootDir: string;
-  let binDir: string;
-
-  beforeEach(async () => {
-    downloadRipGrepMock.mockReset();
-    downloadRipGrepMock.mockResolvedValue(undefined);
-    tempRootDir = await fs.mkdtemp(path.join(os.tmpdir(), 'ripgrep-bin-'));
-    binDir = path.join(tempRootDir, 'bin');
-    await fs.mkdir(binDir, { recursive: true });
-    storageSpy.mockImplementation(() => binDir);
-  });
-
-  afterEach(async () => {
-    storageSpy.mockImplementation(() => originalGetGlobalBinDir());
-    await fs.rm(tempRootDir, { recursive: true, force: true });
+  beforeEach(() => {
+    vi.mocked(fileExists).mockReset();
   });
 
   it('should return true if ripgrep already exists', async () => {
-    const existingPath = path.join(binDir, getRipgrepBinaryName());
-    await fs.writeFile(existingPath, '');
-
+    vi.mocked(fileExists).mockResolvedValue(true);
     const result = await canUseRipgrep();
     expect(result).toBe(true);
-    expect(downloadRipGrepMock).not.toHaveBeenCalled();
   });
 
-  it('should download ripgrep and return true if it does not exist initially', async () => {
-    const expectedPath = path.join(binDir, getRipgrepBinaryName());
-
-    downloadRipGrepMock.mockImplementation(async () => {
-      await fs.writeFile(expectedPath, '');
-    });
-
+  it('should return false if file does not exist', async () => {
+    vi.mocked(fileExists).mockResolvedValue(false);
     const result = await canUseRipgrep();
-
-    expect(result).toBe(true);
-    expect(downloadRipGrep).toHaveBeenCalledWith(binDir);
-    await expect(fs.access(expectedPath)).resolves.toBeUndefined();
-  });
-
-  it('should return false if download fails and file does not exist', async () => {
-    const result = await canUseRipgrep();
-
     expect(result).toBe(false);
-    expect(downloadRipGrep).toHaveBeenCalledWith(binDir);
-  });
-
-  it('should propagate errors from downloadRipGrep', async () => {
-    const error = new Error('Download failed');
-    downloadRipGrepMock.mockRejectedValue(error);
-
-    await expect(canUseRipgrep()).rejects.toThrow(error);
-    expect(downloadRipGrep).toHaveBeenCalledWith(binDir);
-  });
-
-  it('should only download once when called concurrently', async () => {
-    const expectedPath = path.join(binDir, getRipgrepBinaryName());
-
-    downloadRipGrepMock.mockImplementation(
-      () =>
-        new Promise<void>((resolve, reject) => {
-          setTimeout(() => {
-            fs.writeFile(expectedPath, '')
-              .then(() => resolve())
-              .catch(reject);
-          }, 0);
-        }),
-    );
-
-    const firstCall = ensureRgPath();
-    const secondCall = ensureRgPath();
-
-    const [pathOne, pathTwo] = await Promise.all([firstCall, secondCall]);
-
-    expect(pathOne).toBe(expectedPath);
-    expect(pathTwo).toBe(expectedPath);
-    expect(downloadRipGrepMock).toHaveBeenCalledTimes(1);
-    await expect(fs.access(expectedPath)).resolves.toBeUndefined();
   });
 });
 
 describe('ensureRgPath', () => {
-  let tempRootDir: string;
-  let binDir: string;
-
-  beforeEach(async () => {
-    downloadRipGrepMock.mockReset();
-    downloadRipGrepMock.mockResolvedValue(undefined);
-    tempRootDir = await fs.mkdtemp(path.join(os.tmpdir(), 'ripgrep-bin-'));
-    binDir = path.join(tempRootDir, 'bin');
-    await fs.mkdir(binDir, { recursive: true });
-    storageSpy.mockImplementation(() => binDir);
-  });
-
-  afterEach(async () => {
-    storageSpy.mockImplementation(() => originalGetGlobalBinDir());
-    await fs.rm(tempRootDir, { recursive: true, force: true });
+  beforeEach(() => {
+    vi.mocked(fileExists).mockReset();
   });
 
   it('should return rg path if ripgrep already exists', async () => {
-    const existingPath = path.join(binDir, getRipgrepBinaryName());
-    await fs.writeFile(existingPath, '');
-
+    vi.mocked(fileExists).mockResolvedValue(true);
     const rgPath = await ensureRgPath();
-    expect(rgPath).toBe(existingPath);
-    expect(downloadRipGrep).not.toHaveBeenCalled();
+    expect(rgPath).toBe(await getRipgrepPath());
   });
 
-  it('should return rg path if ripgrep is downloaded successfully', async () => {
-    const expectedPath = path.join(binDir, getRipgrepBinaryName());
-
-    downloadRipGrepMock.mockImplementation(async () => {
-      await fs.writeFile(expectedPath, '');
-    });
-
-    const rgPath = await ensureRgPath();
-    expect(rgPath).toBe(expectedPath);
-    expect(downloadRipGrep).toHaveBeenCalledTimes(1);
-    await expect(fs.access(expectedPath)).resolves.toBeUndefined();
+  it('should throw an error if ripgrep cannot be used', async () => {
+    vi.mocked(fileExists).mockResolvedValue(false);
+    await expect(ensureRgPath()).rejects.toThrow(
+      /Cannot find bundled ripgrep binary/,
+    );
   });
-
-  it('should throw an error if ripgrep cannot be used after download attempt', async () => {
-    await expect(ensureRgPath()).rejects.toThrow('Cannot use ripgrep.');
-    expect(downloadRipGrep).toHaveBeenCalledTimes(1);
-  });
-
-  it('should propagate errors from downloadRipGrep', async () => {
-    const error = new Error('Download failed');
-    downloadRipGrepMock.mockRejectedValue(error);
-
-    await expect(ensureRgPath()).rejects.toThrow(error);
-    expect(downloadRipGrep).toHaveBeenCalledWith(binDir);
-  });
-
-  it.runIf(process.platform === 'win32')(
-    'should detect ripgrep when only rg.exe exists on Windows',
-    async () => {
-      const expectedRgExePath = path.join(binDir, 'rg.exe');
-      await fs.writeFile(expectedRgExePath, '');
-
-      const rgPath = await ensureRgPath();
-      expect(rgPath).toBe(expectedRgExePath);
-      expect(downloadRipGrep).not.toHaveBeenCalled();
-      await expect(fs.access(expectedRgExePath)).resolves.toBeUndefined();
-    },
-  );
 });
 
 // Helper function to create mock spawn implementations
@@ -247,9 +124,6 @@ function createMockSpawn(
 
 describe('RipGrepTool', () => {
   let tempRootDir: string;
-  let tempBinRoot: string;
-  let binDir: string;
-  let ripgrepBinaryPath: string;
   let grepTool: RipGrepTool;
   const abortSignal = new AbortController().signal;
 
@@ -266,19 +140,12 @@ describe('RipGrepTool', () => {
   } as unknown as Config;
 
   beforeEach(async () => {
-    downloadRipGrepMock.mockReset();
-    downloadRipGrepMock.mockResolvedValue(undefined);
     mockSpawn.mockReset();
     mockSpawn.mockImplementation(createMockSpawn());
-    tempBinRoot = await fs.mkdtemp(path.join(os.tmpdir(), 'ripgrep-bin-'));
-    binDir = path.join(tempBinRoot, 'bin');
-    await fs.mkdir(binDir, { recursive: true });
-    const binaryName = process.platform === 'win32' ? 'rg.exe' : 'rg';
-    ripgrepBinaryPath = path.join(binDir, binaryName);
-    await fs.writeFile(ripgrepBinaryPath, '');
-    storageSpy.mockImplementation(() => binDir);
     tempRootDir = await fs.mkdtemp(path.join(os.tmpdir(), 'grep-tool-root-'));
 
+    vi.mocked(fileExists).mockResolvedValue(true);
+
     mockConfig = {
       getTargetDir: () => tempRootDir,
       getWorkspaceContext: () => createMockWorkspaceContext(tempRootDir),
@@ -335,9 +202,7 @@ describe('RipGrepTool', () => {
   });
 
   afterEach(async () => {
-    storageSpy.mockImplementation(() => originalGetGlobalBinDir());
     await fs.rm(tempRootDir, { recursive: true, force: true });
-    await fs.rm(tempBinRoot, { recursive: true, force: true });
   });
 
   describe('validateToolParams', () => {
@@ -834,16 +699,16 @@ describe('RipGrepTool', () => {
     });
 
     it('should throw an error if ripgrep is not available', async () => {
-      await fs.rm(ripgrepBinaryPath, { force: true });
-      downloadRipGrepMock.mockResolvedValue(undefined);
+      vi.mocked(fileExists).mockResolvedValue(false);
 
       const params: RipGrepToolParams = { pattern: 'world' };
       const invocation = grepTool.build(params);
 
-      expect(await invocation.execute({ abortSignal })).toStrictEqual({
-        llmContent: 'Error during grep search operation: Cannot use ripgrep.',
-        returnDisplay: 'Error: Cannot use ripgrep.',
-      });
+      const result = await invocation.execute({ abortSignal });
+      expect(result.llmContent).toContain('Cannot find bundled ripgrep binary');
+
+      // restore the mock for subsequent tests
+      vi.mocked(fileExists).mockResolvedValue(true);
     });
   });
 
@@ -2080,6 +1945,68 @@ describe('RipGrepTool', () => {
   });
 });
 
-afterAll(() => {
-  storageSpy.mockRestore();
+describe('getRipgrepPath', () => {
+  afterEach(() => {
+    vi.restoreAllMocks();
+  });
+
+  describe('OS/Architecture Resolution', () => {
+    it.each([
+      { platform: 'darwin', arch: 'arm64', expectedBin: 'rg-darwin-arm64' },
+      { platform: 'darwin', arch: 'x64', expectedBin: 'rg-darwin-x64' },
+      { platform: 'linux', arch: 'arm64', expectedBin: 'rg-linux-arm64' },
+      { platform: 'linux', arch: 'x64', expectedBin: 'rg-linux-x64' },
+      { platform: 'win32', arch: 'x64', expectedBin: 'rg-win32-x64.exe' },
+    ])(
+      'should map $platform $arch to $expectedBin',
+      async ({ platform, arch, expectedBin }) => {
+        vi.spyOn(os, 'platform').mockReturnValue(platform as NodeJS.Platform);
+        vi.spyOn(os, 'arch').mockReturnValue(arch);
+        vi.mocked(fileExists).mockImplementation(async (checkPath) =>
+          checkPath.endsWith(expectedBin),
+        );
+
+        const resolvedPath = await getRipgrepPath();
+        expect(resolvedPath).not.toBeNull();
+        expect(resolvedPath?.endsWith(expectedBin)).toBe(true);
+      },
+    );
+  });
+
+  describe('Path Fallback Logic', () => {
+    beforeEach(() => {
+      vi.spyOn(os, 'platform').mockReturnValue('linux');
+      vi.spyOn(os, 'arch').mockReturnValue('x64');
+    });
+
+    it('should resolve the SEA (flattened) path first', async () => {
+      vi.mocked(fileExists).mockImplementation(async (checkPath) =>
+        checkPath.includes(path.normalize('tools/vendor/ripgrep')),
+      );
+
+      const resolvedPath = await getRipgrepPath();
+      expect(resolvedPath).not.toBeNull();
+      expect(resolvedPath).toContain(path.normalize('tools/vendor/ripgrep'));
+    });
+
+    it('should fall back to the Dev path if SEA path is missing', async () => {
+      vi.mocked(fileExists).mockImplementation(
+        async (checkPath) =>
+          checkPath.includes(path.normalize('core/vendor/ripgrep')) &&
+          !checkPath.includes('tools'),
+      );
+
+      const resolvedPath = await getRipgrepPath();
+      expect(resolvedPath).not.toBeNull();
+      expect(resolvedPath).toContain(path.normalize('core/vendor/ripgrep'));
+      expect(resolvedPath).not.toContain('tools');
+    });
+
+    it('should return null if binary is missing from both paths', async () => {
+      vi.mocked(fileExists).mockResolvedValue(false);
+
+      const resolvedPath = await getRipgrepPath();
+      expect(resolvedPath).toBeNull();
+    });
+  });
 });

packages/core/src/tools/ripGrep.ts

@@ -8,7 +8,8 @@ import type { MessageBus } from '../confirmation-bus/message-bus.js';
 import fs from 'node:fs';
 import fsPromises from 'node:fs/promises';
 import path from 'node:path';
-import { downloadRipGrep } from '@joshua.litt/get-ripgrep';
+import os from 'node:os';
+import { fileURLToPath } from 'node:url';
 import {
   BaseDeclarativeTool,
   BaseToolInvocation,
@@ -22,7 +23,6 @@ import { makeRelative, shortenPath } from '../utils/paths.js';
 import { getErrorMessage, isNodeError } from '../utils/errors.js';
 import type { Config } from '../config/config.js';
 import { fileExists } from '../utils/fileUtils.js';
-import { Storage } from '../config/storage.js';
 import { GREP_TOOL_NAME } from './tool-names.js';
 import { debugLogger } from '../utils/debugLogger.js';
 import {
@@ -39,73 +39,48 @@ import { RIP_GREP_DEFINITION } from './definitions/coreTools.js';
 import { resolveToolDeclaration } from './definitions/resolver.js';
 import { type GrepMatch, formatGrepResults } from './grep-utils.js';
 
-function getRgCandidateFilenames(): readonly string[] {
-  return process.platform === 'win32' ? ['rg.exe', 'rg'] : ['rg'];
-}
+const __dirname = path.dirname(fileURLToPath(import.meta.url));
 
-async function resolveExistingRgPath(): Promise<string | null> {
-  const binDir = Storage.getGlobalBinDir();
-  for (const fileName of getRgCandidateFilenames()) {
-    const candidatePath = path.join(binDir, fileName);
-    if (await fileExists(candidatePath)) {
-      return candidatePath;
+export async function getRipgrepPath(): Promise<string | null> {
+  const platform = os.platform();
+  const arch = os.arch();
+
+  // Map to the correct bundled binary
+  const binName = `rg-${platform}-${arch}${platform === 'win32' ? '.exe' : ''}`;
+
+  const candidatePaths = [
+    // 1. SEA runtime layout: everything is flattened into the root dir
+    path.resolve(__dirname, 'vendor/ripgrep', binName),
+    // 2. Dev/Dist layout: packages/core/dist/tools/ripGrep.js -> packages/core/vendor/ripgrep
+    path.resolve(__dirname, '../../vendor/ripgrep', binName),
+  ];
+
+  for (const candidate of candidatePaths) {
+    if (await fileExists(candidate)) {
+      return candidate;
     }
   }
+
   return null;
 }
 
-let ripgrepAcquisitionPromise: Promise<string | null> | null = null;
 /**
- * Ensures a ripgrep binary is available.
- *
- * NOTE:
- * - The Gemini CLI currently prefers a managed ripgrep binary downloaded
- *   into its global bin directory.
- * - Even if ripgrep is available on the system PATH, it is intentionally
- *   not used at this time.
- *
- * Preference for system-installed ripgrep is blocked on:
- * - checksum verification of external binaries
- * - internalization of the get-ripgrep dependency
- *
- * See:
- * - feat(core): Prefer rg in system path (#11847)
- * - Move get-ripgrep to third_party (#12099)
- */
-async function ensureRipgrepAvailable(): Promise<string | null> {
-  const existingPath = await resolveExistingRgPath();
-  if (existingPath) {
-    return existingPath;
-  }
-  if (!ripgrepAcquisitionPromise) {
-    ripgrepAcquisitionPromise = (async () => {
-      try {
-        await downloadRipGrep(Storage.getGlobalBinDir());
-        return await resolveExistingRgPath();
-      } finally {
-        ripgrepAcquisitionPromise = null;
-      }
-    })();
-  }
-  return ripgrepAcquisitionPromise;
-}
-
-/**
- * Checks if `rg` exists, if not then attempt to download it.
+ * Checks if `rg` exists in the bundled vendor directory.
  */
 export async function canUseRipgrep(): Promise<boolean> {
-  return (await ensureRipgrepAvailable()) !== null;
+  const binPath = await getRipgrepPath();
+  return binPath !== null;
 }
 
 /**
- * Ensures `rg` is downloaded, or throws.
+ * Ensures `rg` is available, or throws.
  */
 export async function ensureRgPath(): Promise<string> {
-  const downloadedPath = await ensureRipgrepAvailable();
-  if (downloadedPath) {
-    return downloadedPath;
+  const binPath = await getRipgrepPath();
+  if (binPath !== null) {
+    return binPath;
   }
-  throw new Error('Cannot use ripgrep.');
+  throw new Error(`Cannot find bundled ripgrep binary.`);
 }
 
 /**

packages/core/src/tools/shell.test.ts

@@ -119,6 +119,7 @@ describe('ShellTool', () => {
       getDebugMode: vi.fn().mockReturnValue(false),
       getTargetDir: vi.fn().mockReturnValue(tempRootDir),
       getSummarizeToolOutputConfig: vi.fn().mockReturnValue(undefined),
+      getTruncateToolOutputThreshold: vi.fn().mockReturnValue(40000),
       getWorkspaceContext: vi
         .fn()
         .mockReturnValue(new WorkspaceContext(tempRootDir)),

packages/core/src/tools/shell.ts

@@ -9,7 +9,7 @@ import fs from 'node:fs';
 import path from 'node:path';
 import os from 'node:os';
 import crypto from 'node:crypto';
-import { debugLogger } from '../index.js';
+import { debugLogger } from '../utils/debugLogger.js';
 import { type SandboxPermissions } from '../services/sandboxManager.js';
 import { ToolErrorType } from './tool-error.js';
 import {
@@ -34,7 +34,10 @@ import {
   type ShellOutputEvent,
 } from '../services/shellExecutionService.js';
 import { formatBytes } from '../utils/formatters.js';
-import type { AnsiOutput } from '../utils/terminalSerializer.js';
+import {
+  serializeAnsiOutputToText,
+  type AnsiOutput,
+} from '../utils/terminalSerializer.js';
 import {
   getCommandRoots,
   initializeShellParsers,
@@ -45,6 +48,7 @@ import {
 } from '../utils/shell-utils.js';
 import { SHELL_TOOL_NAME } from './tool-names.js';
 import { PARAM_ADDITIONAL_PERMISSIONS } from './definitions/base-declarations.js';
+import { ApprovalMode } from '../policy/types.js';
 import type { MessageBus } from '../confirmation-bus/message-bus.js';
 import { getShellDefinition } from './definitions/coreTools.js';
 import { resolveToolDeclaration } from './definitions/resolver.js';
@@ -249,6 +253,10 @@ export class ShellToolInvocation extends BaseToolInvocation<
     abortSignal: AbortSignal,
     forcedDecision?: ForcedToolDecision,
   ): Promise<ToolCallConfirmationDetails | false> {
+    if (this.context.config.getApprovalMode() === ApprovalMode.YOLO) {
+      return super.shouldConfirmExecute(abortSignal, forcedDecision);
+    }
+
     if (this.params[PARAM_ADDITIONAL_PERMISSIONS]) {
       return this.getConfirmationDetails(abortSignal);
     }
@@ -459,6 +467,14 @@ export class ShellToolInvocation extends BaseToolInvocation<
 
     const onAbort = () => combinedController.abort();
 
+    const outputFileName = `gemini_shell_output_${crypto.randomBytes(6).toString('hex')}.log`;
+    const projectTempDir = this.context.config.storage.getProjectTempDir();
+    fs.mkdirSync(projectTempDir, { recursive: true });
+    const outputFilePath = path.join(projectTempDir, outputFileName);
+    const outputStream = fs.createWriteStream(outputFilePath);
+
+    let fullOutputReturned = false;
+
     try {
       // pgrep is not available on Windows, so we can't get background PIDs
       const commandToExecute = this.wrapCommandForPgrep(
@@ -485,6 +501,7 @@ export class ShellToolInvocation extends BaseToolInvocation<
       let cumulativeOutput: string | AnsiOutput = '';
       let lastUpdateTime = Date.now();
       let isBinaryStream = false;
+      let totalBytesWritten = 0;
 
       const resetTimeout = () => {
         if (timeoutMs <= 0) {
@@ -510,31 +527,65 @@ export class ShellToolInvocation extends BaseToolInvocation<
           cwd,
           (event: ShellOutputEvent) => {
             resetTimeout(); // Reset timeout on any event
-            if (!updateOutput) {
-              return;
-            }
-
-            let shouldUpdate = false;
 
             switch (event.type) {
+              case 'raw_data':
+                // We do not write raw data to the file to avoid spurious escape codes.
+                // We rely on 'file_data' for the clean output stream.
+                break;
+              case 'file_data':
+                totalBytesWritten += Buffer.byteLength(event.chunk);
+                outputStream.write(event.chunk);
+                break;
               case 'data':
                 if (isBinaryStream) break;
-                cumulativeOutput = event.chunk;
-                shouldUpdate = true;
+                if (typeof event.chunk === 'string') {
+                  if (typeof cumulativeOutput === 'string') {
+                    // Accumulate string chunks, capped at 16MB (same as ShellExecutionService)
+                    const MAX_BUFFER = 16 * 1024 * 1024;
+                    if (
+                      cumulativeOutput.length + event.chunk.length >
+                      MAX_BUFFER
+                    ) {
+                      cumulativeOutput = (cumulativeOutput + event.chunk).slice(
+                        -MAX_BUFFER,
+                      );
+                    } else {
+                      cumulativeOutput += event.chunk;
+                    }
+                  } else {
+                    // In case of mode switch (unlikely)
+                    cumulativeOutput = event.chunk;
+                  }
+                } else {
+                  // Snapshots (AnsiOutput) always replace
+                  cumulativeOutput = event.chunk;
+                }
+                if (updateOutput && !this.params.is_background) {
+                  updateOutput(cumulativeOutput);
+                  lastUpdateTime = Date.now();
+                }
                 break;
               case 'binary_detected':
                 isBinaryStream = true;
                 cumulativeOutput =
                   '[Binary output detected. Halting stream...]';
-                shouldUpdate = true;
+                if (updateOutput && !this.params.is_background) {
+                  updateOutput(cumulativeOutput);
+                }
                 break;
               case 'binary_progress':
                 isBinaryStream = true;
                 cumulativeOutput = `[Receiving binary output... ${formatBytes(
                   event.bytesReceived,
                 )} received]`;
-                if (Date.now() - lastUpdateTime > OUTPUT_UPDATE_INTERVAL_MS) {
-                  shouldUpdate = true;
+                if (
+                  updateOutput &&
+                  !this.params.is_background &&
+                  Date.now() - lastUpdateTime > OUTPUT_UPDATE_INTERVAL_MS
+                ) {
+                  updateOutput(cumulativeOutput);
+                  lastUpdateTime = Date.now();
                 }
                 break;
               case 'exit':
@@ -543,11 +594,6 @@ export class ShellToolInvocation extends BaseToolInvocation<
                 throw new Error('An unhandled ShellOutputEvent was found.');
               }
             }
-
-            if (shouldUpdate && !this.params.is_background) {
-              updateOutput(cumulativeOutput);
-              lastUpdateTime = Date.now();
-            }
           },
           combinedController.signal,
           this.context.config.getEnableInteractiveShell(),
@@ -610,9 +656,14 @@ export class ShellToolInvocation extends BaseToolInvocation<
           await new Promise((resolve) => setTimeout(resolve, delay));
 
           if (!completed) {
+            const initialOutputText =
+              typeof cumulativeOutput === 'string'
+                ? cumulativeOutput
+                : serializeAnsiOutputToText(cumulativeOutput);
+
             // Return early with initial output if still running
             return {
-              llmContent: `Command is running in background. PID: ${pid}. Initial output:\n${cumulativeOutput}`,
+              llmContent: `Command is running in background. PID: ${pid}. Initial output:\n${initialOutputText}`,
               returnDisplay: `Background process started with PID ${pid}.`,
             };
           }
@@ -620,6 +671,15 @@ export class ShellToolInvocation extends BaseToolInvocation<
       }
 
       const result = await resultPromise;
+      await new Promise<void>((resolve, reject) => {
+        outputStream.on('error', reject);
+        outputStream.end(resolve);
+      });
+
+      // Ensure the stream is fully closed before we proceed
+      if (!outputStream.closed) {
+        await new Promise<void>((resolve) => outputStream.on('close', resolve));
+      }
 
       const backgroundPIDs: number[] = [];
       if (os.platform() !== 'win32') {
@@ -913,21 +973,46 @@ export class ShellToolInvocation extends BaseToolInvocation<
           this.context.geminiClient,
           signal,
         );
-        return {
+        const threshold = this.context.config.getTruncateToolOutputThreshold();
+        const fullOutputFilePath =
+          threshold > 0 && totalBytesWritten >= threshold
+            ? outputFilePath
+            : undefined;
+
+        const toolResult: ToolResult = {
           llmContent: summary,
           returnDisplay,
+          fullOutputFilePath,
           ...executionError,
         };
+        if (toolResult.fullOutputFilePath) {
+          fullOutputReturned = true;
+        }
+        return toolResult;
       }
 
-      return {
+      const threshold = this.context.config.getTruncateToolOutputThreshold();
+      const fullOutputFilePath =
+        threshold > 0 && totalBytesWritten >= threshold
+          ? outputFilePath
+          : undefined;
+
+      const toolResult: ToolResult = {
         llmContent,
         returnDisplay,
         data,
+        fullOutputFilePath,
         ...executionError,
       };
+      if (toolResult.fullOutputFilePath) {
+        fullOutputReturned = true;
+      }
+      return toolResult;
     } finally {
       if (timeoutTimer) clearTimeout(timeoutTimer);
+      if (!outputStream.closed) {
+        outputStream.destroy();
+      }
       signal.removeEventListener('abort', onAbort);
       timeoutController.signal.removeEventListener('abort', onAbort);
       try {
@@ -935,6 +1020,13 @@ export class ShellToolInvocation extends BaseToolInvocation<
       } catch {
         // Ignore errors during unlink
       }
+      if (!fullOutputReturned) {
+        try {
+          await fsPromises.unlink(outputFilePath);
+        } catch {
+          // Ignore errors during unlink
+        }
+      }
     }
   }
 }

packages/core/src/tools/tools.ts

@@ -776,6 +776,13 @@ export interface ToolResult {
     name: string;
     args: Record<string, unknown>;
   };
+
+  /**
+   * Optional path to a file containing the full, non-truncated output of the tool.
+   * If provided, the scheduler may use this file for long-term storage and
+   * reference it in the conversation history if the output is truncated.
+   */
+  fullOutputFilePath?: string;
 }
 
 /**

packages/core/src/utils/compatibility.test.ts

@@ -195,6 +195,7 @@ describe('compatibility', () => {
         desc: '256 colors are not supported',
       },
     ])('should return $expected when $desc', ({ depth, term, expected }) => {
+      vi.stubEnv('COLORTERM', '');
       process.stdout.getColorDepth = vi.fn().mockReturnValue(depth);
       if (term !== undefined) {
         vi.stubEnv('TERM', term);
@@ -203,6 +204,13 @@ describe('compatibility', () => {
       }
       expect(supports256Colors()).toBe(expected);
     });
+
+    it('should return true when COLORTERM is kmscon', () => {
+      process.stdout.getColorDepth = vi.fn().mockReturnValue(4);
+      vi.stubEnv('TERM', 'linux');
+      vi.stubEnv('COLORTERM', 'kmscon');
+      expect(supports256Colors()).toBe(true);
+    });
   });
 
   describe('supportsTrueColor', () => {
@@ -230,6 +238,12 @@ describe('compatibility', () => {
         expected: true,
         desc: 'getColorDepth returns >= 24',
       },
+      {
+        colorterm: 'kmscon',
+        depth: 4,
+        expected: true,
+        desc: 'COLORTERM is kmscon',
+      },
       {
         colorterm: '',
         depth: 8,
@@ -409,6 +423,18 @@ describe('compatibility', () => {
       );
     });
 
+    it('should return no color warnings for kmscon terminal', () => {
+      vi.mocked(os.platform).mockReturnValue('linux');
+      vi.stubEnv('TERMINAL_EMULATOR', '');
+      vi.stubEnv('TERM', 'linux');
+      vi.stubEnv('COLORTERM', 'kmscon');
+      process.stdout.getColorDepth = vi.fn().mockReturnValue(4);
+
+      const warnings = getCompatibilityWarnings();
+      expect(warnings.find((w) => w.id === '256-color')).toBeUndefined();
+      expect(warnings.find((w) => w.id === 'true-color')).toBeUndefined();
+    });
+
     it('should return no warnings in a standard environment with true color', () => {
       vi.mocked(os.platform).mockReturnValue('darwin');
       vi.stubEnv('TERMINAL_EMULATOR', '');

packages/core/src/utils/compatibility.ts

@@ -85,6 +85,11 @@ export function supports256Colors(): boolean {
     return true;
   }
 
+  // Terminals supporting true color (like kmscon) also support 256 colors
+  if (supportsTrueColor()) {
+    return true;
+  }
+
   return false;
 }
 
@@ -95,7 +100,8 @@ export function supportsTrueColor(): boolean {
   // Check COLORTERM environment variable
   if (
     process.env['COLORTERM'] === 'truecolor' ||
-    process.env['COLORTERM'] === '24bit'
+    process.env['COLORTERM'] === '24bit' ||
+    process.env['COLORTERM'] === 'kmscon'
   ) {
     return true;
   }

packages/core/src/utils/fileUtils.ts

@@ -677,6 +677,48 @@ ${head}
 ${tail}`;
 }
 
+/**
+ * Moves tool output from a source path to a temporary file for later retrieval.
+ */
+export async function moveToolOutputToFile(
+  sourcePath: string,
+  toolName: string,
+  id: string | number, // Accept string (callId) or number (truncationId)
+  projectTempDir: string,
+  sessionId?: string,
+): Promise<{ outputFile: string }> {
+  const safeToolName = sanitizeFilenamePart(toolName).toLowerCase();
+  const safeId = sanitizeFilenamePart(id.toString()).toLowerCase();
+  const fileName = safeId.startsWith(safeToolName)
+    ? `${safeId}.txt`
+    : `${safeToolName}_${safeId}.txt`;
+
+  let toolOutputDir = path.join(projectTempDir, TOOL_OUTPUTS_DIR);
+  if (sessionId) {
+    const safeSessionId = sanitizeFilenamePart(sessionId);
+    toolOutputDir = path.join(toolOutputDir, `session-${safeSessionId}`);
+  }
+  const outputFile = path.join(toolOutputDir, fileName);
+
+  await fsPromises.mkdir(toolOutputDir, { recursive: true });
+
+  try {
+    // Attempt rename (efficient if on the same filesystem)
+    await fsPromises.rename(sourcePath, outputFile);
+  } catch (error: unknown) {
+    // If rename fails (e.g. cross-filesystem), copy and then delete
+    // eslint-disable-next-line @typescript-eslint/no-unsafe-type-assertion
+    if ((error as { code?: string }).code === 'EXDEV') {
+      await fsPromises.copyFile(sourcePath, outputFile);
+      await fsPromises.unlink(sourcePath);
+    } else {
+      throw error;
+    }
+  }
+
+  return { outputFile };
+}
+
 /**
  * Saves tool output to a temporary file for later retrieval.
  */

packages/core/src/utils/generateContentResponseUtilities.ts

@@ -22,12 +22,13 @@ function createFunctionResponsePart(
   callId: string,
   toolName: string,
   output: string,
+  outputFile?: string,
 ): Part {
   return {
     functionResponse: {
       id: callId,
       name: toolName,
-      response: { output },
+      response: { output, outputFile },
     },
   };
 }
@@ -50,9 +51,12 @@ export function convertToFunctionResponse(
   llmContent: PartListUnion,
   model: string,
   config?: Config,
+  outputFile?: string,
 ): Part[] {
   if (typeof llmContent === 'string') {
-    return [createFunctionResponsePart(callId, toolName, llmContent)];
+    return [
+      createFunctionResponsePart(callId, toolName, llmContent, outputFile),
+    ];
   }
 
   const parts = toParts(llmContent);
@@ -94,7 +98,10 @@ export function convertToFunctionResponse(
     functionResponse: {
       id: callId,
       name: toolName,
-      response: textParts.length > 0 ? { output: textParts.join('\n') } : {},
+      response: {
+        ...(textParts.length > 0 ? { output: textParts.join('\n') } : {}),
+        outputFile,
+      },
     },
   };
 

packages/core/src/utils/retry.test.ts

@@ -511,6 +511,40 @@ describe('retryWithBackoff', () => {
       expect(mockFn).toHaveBeenCalledTimes(2);
     });
 
+    it('should retry on OpenSSL 3.x SSL error code (ERR_SSL_SSL/TLS_ALERT_BAD_RECORD_MAC)', async () => {
+      const error = new Error('SSL error');
+      (error as any).code = 'ERR_SSL_SSL/TLS_ALERT_BAD_RECORD_MAC';
+      const mockFn = vi
+        .fn()
+        .mockRejectedValueOnce(error)
+        .mockResolvedValue('success');
+
+      const promise = retryWithBackoff(mockFn, {
+        initialDelayMs: 1,
+        maxDelayMs: 1,
+      });
+      await vi.runAllTimersAsync();
+      await expect(promise).resolves.toBe('success');
+      expect(mockFn).toHaveBeenCalledTimes(2);
+    });
+
+    it('should retry on unknown SSL BAD_RECORD_MAC variant via substring fallback', async () => {
+      const error = new Error('SSL error');
+      (error as any).code = 'ERR_SSL_SOME_FUTURE_BAD_RECORD_MAC';
+      const mockFn = vi
+        .fn()
+        .mockRejectedValueOnce(error)
+        .mockResolvedValue('success');
+
+      const promise = retryWithBackoff(mockFn, {
+        initialDelayMs: 1,
+        maxDelayMs: 1,
+      });
+      await vi.runAllTimersAsync();
+      await expect(promise).resolves.toBe('success');
+      expect(mockFn).toHaveBeenCalledTimes(2);
+    });
+
     it('should retry on gaxios-style SSL error with code property', async () => {
       // This matches the exact structure from issue #17318
       const error = new Error(

packages/core/src/utils/retry.ts

@@ -53,14 +53,30 @@ const RETRYABLE_NETWORK_CODES = [
   'ENOTFOUND',
   'EAI_AGAIN',
   'ECONNREFUSED',
-  // SSL/TLS transient errors
-  'ERR_SSL_SSLV3_ALERT_BAD_RECORD_MAC',
   'ERR_SSL_WRONG_VERSION_NUMBER',
-  'ERR_SSL_DECRYPTION_FAILED_OR_BAD_RECORD_MAC',
-  'ERR_SSL_BAD_RECORD_MAC',
   'EPROTO', // Generic protocol error (often SSL-related)
 ];
 
+// Node.js builds SSL error codes by prepending ERR_SSL_ to the uppercased
+// OpenSSL reason string with spaces replaced by underscores (see
+// TLSWrap::ClearOut in node/src/crypto/crypto_tls.cc). The reason string
+// format varies by OpenSSL version (e.g. ERR_SSL_SSLV3_ALERT_BAD_RECORD_MAC
+// on OpenSSL 1.x, ERR_SSL_SSL/TLS_ALERT_BAD_RECORD_MAC on OpenSSL 3.x), so
+// match the stable suffix instead of enumerating every variant.
+const RETRYABLE_SSL_ERROR_PATTERN = /^ERR_SSL_.*BAD_RECORD_MAC/i;
+
+/**
+ * Returns true if the error code should be retried: either an exact match
+ * against RETRYABLE_NETWORK_CODES, or an SSL BAD_RECORD_MAC variant (the
+ * OpenSSL reason-string portion of the code varies across OpenSSL versions).
+ */
+function isRetryableSslErrorCode(code: string): boolean {
+  return (
+    RETRYABLE_NETWORK_CODES.includes(code) ||
+    RETRYABLE_SSL_ERROR_PATTERN.test(code)
+  );
+}
+
 function getNetworkErrorCode(error: unknown): string | undefined {
   const getCode = (obj: unknown): string | undefined => {
     if (typeof obj !== 'object' || obj === null) {
@@ -112,7 +128,7 @@ export function getRetryErrorType(error: unknown): string {
   }
 
   const errorCode = getNetworkErrorCode(error);
-  if (errorCode && RETRYABLE_NETWORK_CODES.includes(errorCode)) {
+  if (errorCode && isRetryableSslErrorCode(errorCode)) {
     return errorCode;
   }
 
@@ -153,7 +169,7 @@ export function isRetryableError(
 ): boolean {
   // Check for common network error codes
   const errorCode = getNetworkErrorCode(error);
-  if (errorCode && RETRYABLE_NETWORK_CODES.includes(errorCode)) {
+  if (errorCode && isRetryableSslErrorCode(errorCode)) {
     return true;
   }
 

packages/core/src/utils/terminalSerializer.ts

@@ -250,6 +250,12 @@ export function serializeTerminalToObject(
   return result;
 }
 
+export function serializeAnsiOutputToText(output: AnsiOutput): string {
+  return output
+    .map((line) => line.map((segment) => segment.text).join(''))
+    .join('\n');
+}
+
 // ANSI color palette from https://en.wikipedia.org/wiki/ANSI_escape_code#8-bit
 const ANSI_COLORS = [
   '#000000',

packages/devtools/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@google/gemini-cli-devtools",
-  "version": "0.39.0-nightly.20260408.e77b22e63",
+  "version": "0.40.0-nightly.20260414.g5b1f7375a",
   "license": "Apache-2.0",
   "type": "module",
   "main": "dist/src/index.js",

packages/sdk/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@google/gemini-cli-sdk",
-  "version": "0.39.0-nightly.20260408.e77b22e63",
+  "version": "0.40.0-nightly.20260414.g5b1f7375a",
   "description": "Gemini CLI SDK",
   "license": "Apache-2.0",
   "repository": {

packages/test-utils/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@google/gemini-cli-test-utils",
-  "version": "0.39.0-nightly.20260408.e77b22e63",
+  "version": "0.40.0-nightly.20260414.g5b1f7375a",
   "private": true,
   "main": "src/index.ts",
   "license": "Apache-2.0",

packages/test-utils/src/perf-test-harness.ts

@@ -147,7 +147,9 @@ export class PerfTestHarness {
       throw new Error(`No active timer found for label "${label}"`);
     }
 
-    const wallClockMs = performance.now() - timer.startTime;
+    // Round wall-clock time to nearest 0.1 ms
+    const wallClockMs =
+      Math.round((performance.now() - timer.startTime) * 10) / 10;
     const cpuDelta = process.cpuUsage(timer.startCpuUsage);
     this.activeTimers.delete(label);
 

packages/test-utils/src/test-rig.ts

@@ -193,6 +193,28 @@ export function checkModelOutputContent(
   return isValid;
 }
 
+export interface MetricDataPoint {
+  attributes?: Record<string, unknown>;
+  value?: {
+    sum?: number;
+    min?: number;
+    max?: number;
+    count?: number;
+  };
+  startTime?: [number, number];
+  endTime?: string;
+}
+
+export interface TelemetryMetric {
+  descriptor: {
+    name: string;
+    type?: string;
+    description?: string;
+    unit?: string;
+  };
+  dataPoints: MetricDataPoint[];
+}
+
 export interface ParsedLog {
   attributes?: {
     'event.name'?: string;
@@ -213,11 +235,7 @@ export interface ParsedLog {
     prompt_id?: string;
   };
   scopeMetrics?: {
-    metrics: {
-      descriptor: {
-        name: string;
-      };
-    }[];
+    metrics: TelemetryMetric[];
   }[];
 }
 
@@ -1297,6 +1315,10 @@ export class TestRig {
     return logs;
   }
 
+  readTelemetryLogs(): ParsedLog[] {
+    return this._readAndParseTelemetryLog();
+  }
+
   private _readAndParseTelemetryLog(): ParsedLog[] {
     // Telemetry is always written to the test directory
     const logFilePath = join(this.homeDir!, 'telemetry.log');
@@ -1450,7 +1472,7 @@ export class TestRig {
     );
   }
 
-  readMetric(metricName: string): Record<string, unknown> | null {
+  readMetric(metricName: string): TelemetryMetric | null {
     const logs = this._readAndParseTelemetryLog();
     for (const logData of logs) {
       if (logData.scopeMetrics) {

packages/vscode-ide-companion/package.json

@@ -2,7 +2,7 @@
   "name": "gemini-cli-vscode-ide-companion",
   "displayName": "Gemini CLI Companion",
   "description": "Enable Gemini CLI with direct access to your IDE workspace.",
-  "version": "0.39.0-nightly.20260408.e77b22e63",
+  "version": "0.40.0-nightly.20260414.g5b1f7375a",
   "publisher": "google",
   "icon": "assets/icon.png",
   "repository": {

perf-tests/baselines.json

@@ -1,19 +1,24 @@
 {
   "version": 1,
-  "updatedAt": "2026-04-09T02:30:22.000Z",
+  "updatedAt": "2026-04-14T14:04:02.662Z",
   "scenarios": {
     "cold-startup-time": {
-      "wallClockMs": 927.553249999999,
+      "wallClockMs": 927.6,
       "cpuTotalUs": 1470,
       "timestamp": "2026-04-08T22:27:54.871Z"
     },
     "idle-cpu-usage": {
-      "wallClockMs": 5000.460750000002,
+      "wallClockMs": 5000.5,
       "cpuTotalUs": 12157,
       "timestamp": "2026-04-08T22:28:19.098Z"
     },
+    "asian-language-conv": {
+      "wallClockMs": 2315.1,
+      "cpuTotalUs": 6283,
+      "timestamp": "2026-04-14T15:22:56.133Z"
+    },
     "skill-loading-time": {
-      "wallClockMs": 930.0920409999962,
+      "wallClockMs": 930.1,
       "cpuTotalUs": 1323,
       "timestamp": "2026-04-08T22:28:23.290Z"
     },
@@ -21,6 +26,31 @@
       "wallClockMs": 1119.9,
       "cpuTotalUs": 2100,
       "timestamp": "2026-04-09T02:30:22.000Z"
+    },
+    "long-conversation-resume": {
+      "wallClockMs": 4212.5,
+      "cpuTotalUs": 351393,
+      "timestamp": "2026-04-14T14:02:53.268Z"
+    },
+    "long-conversation-typing": {
+      "wallClockMs": 113.7,
+      "cpuTotalUs": 3304,
+      "timestamp": "2026-04-14T14:03:12.525Z"
+    },
+    "long-conversation-execution": {
+      "wallClockMs": 248.7,
+      "cpuTotalUs": 3825,
+      "timestamp": "2026-04-14T14:03:28.575Z"
+    },
+    "long-conversation-terminal-scrolling": {
+      "wallClockMs": 362.4,
+      "cpuTotalUs": 12755860,
+      "timestamp": "2026-04-14T14:03:45.687Z"
+    },
+    "long-conversation-alternate-scrolling": {
+      "wallClockMs": 362.4,
+      "cpuTotalUs": 12755860,
+      "timestamp": "2026-04-14T14:04:02.662Z"
     }
   }
 }

perf-tests/perf-usage.test.ts

@@ -5,10 +5,20 @@
  */
 
 import { describe, it, beforeAll, afterAll } from 'vitest';
-import { TestRig, PerfTestHarness } from '@google/gemini-cli-test-utils';
+import {
+  TestRig,
+  PerfTestHarness,
+  type PerfSnapshot,
+} from '@google/gemini-cli-test-utils';
 import { join, dirname } from 'node:path';
 import { fileURLToPath } from 'node:url';
-import { existsSync, readFileSync } from 'node:fs';
+import {
+  existsSync,
+  readFileSync,
+  mkdirSync,
+  copyFileSync,
+  writeFileSync,
+} from 'node:fs';
 
 const __dirname = dirname(fileURLToPath(import.meta.url));
 const BASELINES_PATH = join(__dirname, 'baselines.json');
@@ -98,6 +108,36 @@ describe('CPU Performance Tests', () => {
     }
   });
 
+  it('asian-language-conv: verify perf is acceptable ', async () => {
+    const result = await harness.runScenario(
+      'asian-language-conv',
+      async () => {
+        const rig = new TestRig();
+        try {
+          rig.setup('perf-asian-language', {
+            fakeResponsesPath: join(__dirname, 'perf.asian-language.responses'),
+          });
+
+          return await harness.measure('asian-language', async () => {
+            await rig.run({
+              args: ['嗨'],
+              timeout: 120000,
+              env: { GEMINI_API_KEY: 'fake-perf-test-key' },
+            });
+          });
+        } finally {
+          await rig.cleanup();
+        }
+      },
+    );
+
+    if (UPDATE_BASELINES) {
+      harness.updateScenarioBaseline(result);
+    } else {
+      harness.assertWithinBaseline(result);
+    }
+  });
+
   it('skill-loading-time: startup with many skills within baseline', async () => {
     const SKILL_COUNT = 20;
 
@@ -165,7 +205,7 @@ describe('CPU Performance Tests', () => {
           const snapshot = await harness.measureWithEventLoop(
             'high-volume-output',
             async () => {
-              const runResult = await rig.run({
+              await rig.run({
                 args: ['Generate 1M lines of output'],
                 timeout: 120000,
                 env: {
@@ -176,7 +216,6 @@ describe('CPU Performance Tests', () => {
                   DEBUG: 'true',
                 },
               });
-              console.log(`  Child Process Output:`, runResult);
             },
           );
 
@@ -216,8 +255,7 @@ describe('CPU Performance Tests', () => {
               JSON.stringify(toolLatencyMetric),
             );
           }
-          // eslint-disable-next-line @typescript-eslint/no-explicit-any
-          const logs = (rig as any)._readAndParseTelemetryLog();
+          const logs = rig.readTelemetryLogs();
           console.log(`  Total telemetry log entries: ${logs.length}`);
           for (const logData of logs) {
             if (logData.scopeMetrics) {
@@ -242,10 +280,9 @@ describe('CPU Performance Tests', () => {
 
             const findValue = (percentile: string) => {
               const dp = eventLoopMetric.dataPoints.find(
-                // eslint-disable-next-line @typescript-eslint/no-explicit-any
-                (p: any) => p.attributes.percentile === percentile,
+                (p) => p.attributes?.['percentile'] === percentile,
               );
-              return dp ? dp.value.min : undefined;
+              return dp?.value?.min;
             };
 
             snapshot.childEventLoopDelayP50Ms = findValue('p50');
@@ -266,4 +303,358 @@ describe('CPU Performance Tests', () => {
       harness.assertWithinBaseline(result);
     }
   });
+
+  describe('long-conversation', () => {
+    let rig: TestRig;
+    const identifier = 'perf-long-conversation';
+    const SESSION_ID =
+      'anonymous_unique_id_577296e0eee5afecdcec05d11838e0cd1a851cd97a28119a4a876b11';
+    const LARGE_CHAT_SOURCE = join(
+      __dirname,
+      '..',
+      'memory-tests',
+      'large-chat-session.json',
+    );
+
+    beforeAll(async () => {
+      if (!existsSync(LARGE_CHAT_SOURCE)) {
+        throw new Error(
+          `Performance test fixture missing: ${LARGE_CHAT_SOURCE}.`,
+        );
+      }
+
+      rig = new TestRig();
+      rig.setup(identifier, {
+        fakeResponsesPath: join(__dirname, 'perf.long-chat.responses'),
+      });
+
+      const geminiDir = join(rig.homeDir!, '.gemini');
+      const projectTempDir = join(geminiDir, 'tmp', identifier);
+      const targetChatsDir = join(projectTempDir, 'chats');
+
+      mkdirSync(targetChatsDir, { recursive: true });
+      writeFileSync(
+        join(geminiDir, 'projects.json'),
+        JSON.stringify({
+          projects: { [rig.testDir!]: identifier },
+        }),
+      );
+      writeFileSync(join(projectTempDir, '.project_root'), rig.testDir!);
+      copyFileSync(
+        LARGE_CHAT_SOURCE,
+        join(targetChatsDir, `session-${SESSION_ID}.json`),
+      );
+    });
+
+    afterAll(async () => {
+      await rig.cleanup();
+    });
+
+    it('session-load: resume a 60MB chat history', async () => {
+      const result = await harness.runScenario(
+        'long-conversation-resume',
+        async () => {
+          const snapshot = await harness.measureWithEventLoop(
+            'resume',
+            async () => {
+              const run = await rig.runInteractive({
+                args: ['--resume', 'latest'],
+                env: {
+                  GEMINI_API_KEY: 'fake-perf-test-key',
+                  GEMINI_TELEMETRY_ENABLED: 'true',
+                  GEMINI_MEMORY_MONITOR_INTERVAL: '500',
+                  GEMINI_EVENT_LOOP_MONITOR_ENABLED: 'true',
+                  DEBUG: 'true',
+                },
+              });
+              await run.kill();
+            },
+          );
+          return snapshot;
+        },
+      );
+
+      if (UPDATE_BASELINES) {
+        harness.updateScenarioBaseline(result);
+      } else {
+        harness.assertWithinBaseline(result);
+      }
+    });
+
+    it('typing: latency when typing into a large session', async () => {
+      const result = await harness.runScenario(
+        'long-conversation-typing',
+        async () => {
+          const run = await rig.runInteractive({
+            args: ['--resume', 'latest'],
+            env: {
+              GEMINI_API_KEY: 'fake-perf-test-key',
+              GEMINI_TELEMETRY_ENABLED: 'true',
+              GEMINI_MEMORY_MONITOR_INTERVAL: '500',
+              GEMINI_EVENT_LOOP_MONITOR_ENABLED: 'true',
+              DEBUG: 'true',
+            },
+          });
+
+          const snapshot = await harness.measureWithEventLoop(
+            'typing',
+            async () => {
+              // On average, the expected latency per key is under 30ms.
+              for (const char of 'Hello') {
+                await run.type(char);
+              }
+            },
+          );
+
+          await run.kill();
+          return snapshot;
+        },
+      );
+
+      if (UPDATE_BASELINES) {
+        harness.updateScenarioBaseline(result);
+      } else {
+        harness.assertWithinBaseline(result);
+      }
+    });
+
+    it('execution: response latency for a simple shell command', async () => {
+      const result = await harness.runScenario(
+        'long-conversation-execution',
+        async () => {
+          const run = await rig.runInteractive({
+            args: ['--resume', 'latest'],
+            env: {
+              GEMINI_API_KEY: 'fake-perf-test-key',
+              GEMINI_TELEMETRY_ENABLED: 'true',
+              GEMINI_MEMORY_MONITOR_INTERVAL: '500',
+              GEMINI_EVENT_LOOP_MONITOR_ENABLED: 'true',
+              DEBUG: 'true',
+            },
+          });
+
+          await run.expectText('Type your message');
+
+          const snapshot = await harness.measureWithEventLoop(
+            'execution',
+            async () => {
+              await run.sendKeys('!echo hi\r');
+              await run.expectText('hi');
+            },
+          );
+
+          await run.kill();
+          return snapshot;
+        },
+      );
+
+      if (UPDATE_BASELINES) {
+        harness.updateScenarioBaseline(result);
+      } else {
+        harness.assertWithinBaseline(result);
+      }
+    });
+
+    it('terminal-scrolling: latency when scrolling a large terminal buffer', async () => {
+      const result = await harness.runScenario(
+        'long-conversation-terminal-scrolling',
+        async () => {
+          // Enable terminalBuffer to intentionally test CLI scrolling logic
+          const settingsPath = join(rig.homeDir!, '.gemini', 'settings.json');
+          writeFileSync(
+            settingsPath,
+            JSON.stringify({
+              security: { folderTrust: { enabled: false } },
+              ui: { terminalBuffer: true },
+            }),
+          );
+
+          const run = await rig.runInteractive({
+            args: ['--resume', 'latest'],
+            env: {
+              GEMINI_API_KEY: 'fake-perf-test-key',
+              GEMINI_TELEMETRY_ENABLED: 'true',
+              GEMINI_MEMORY_MONITOR_INTERVAL: '500',
+              GEMINI_EVENT_LOOP_MONITOR_ENABLED: 'true',
+              DEBUG: 'true',
+            },
+          });
+
+          await run.expectText('Type your message');
+
+          for (let i = 0; i < 5; i++) {
+            await run.sendKeys('\u001b[5~'); // PageUp
+          }
+
+          // Scroll to the very top
+          await run.sendKeys('\u001b[H'); // Home
+          // Verify top line of chat is visible.
+          await run.expectText('Authenticated with');
+
+          for (let i = 0; i < 5; i++) {
+            await run.sendKeys('\u001b[6~'); // PageDown
+          }
+
+          await rig.waitForTelemetryReady();
+          await run.kill();
+
+          const eventLoopMetric = rig.readMetric('event_loop.delay');
+          const cpuMetric = rig.readMetric('cpu.usage');
+
+          let p50Ms = 0;
+          let p95Ms = 0;
+          let maxMs = 0;
+          if (eventLoopMetric) {
+            const dataPoints = eventLoopMetric.dataPoints;
+            const p50Data = dataPoints.find(
+              (dp) => dp.attributes?.['percentile'] === 'p50',
+            );
+            const p95Data = dataPoints.find(
+              (dp) => dp.attributes?.['percentile'] === 'p95',
+            );
+            const maxData = dataPoints.find(
+              (dp) => dp.attributes?.['percentile'] === 'max',
+            );
+
+            if (p50Data?.value?.sum) p50Ms = p50Data.value.sum;
+            if (p95Data?.value?.sum) p95Ms = p95Data.value.sum;
+            if (maxData?.value?.sum) maxMs = maxData.value.sum;
+          }
+
+          let cpuTotalUs = 0;
+          if (cpuMetric) {
+            const dataPoints = cpuMetric.dataPoints;
+            for (const dp of dataPoints) {
+              if (dp.value?.sum && dp.value.sum > 0) {
+                cpuTotalUs += dp.value.sum;
+              }
+            }
+          }
+          const cpuUserUs = cpuTotalUs;
+          const cpuSystemUs = 0;
+
+          const snapshot: PerfSnapshot = {
+            timestamp: Date.now(),
+            label: 'scrolling',
+            wallClockMs: Math.round(p50Ms * 10) / 10,
+            cpuTotalUs,
+            cpuUserUs,
+            cpuSystemUs,
+            eventLoopDelayP50Ms: p50Ms,
+            eventLoopDelayP95Ms: p95Ms,
+            eventLoopDelayMaxMs: maxMs,
+          };
+
+          return snapshot;
+        },
+      );
+
+      if (UPDATE_BASELINES) {
+        harness.updateScenarioBaseline(result);
+      } else {
+        harness.assertWithinBaseline(result);
+      }
+    });
+
+    it('alternate-scrolling: latency when scrolling a large alternate buffer', async () => {
+      const result = await harness.runScenario(
+        'long-conversation-alternate-scrolling',
+        async () => {
+          // Enable useAlternateBuffer to intentionally test CLI scrolling logic
+          const settingsPath = join(rig.homeDir!, '.gemini', 'settings.json');
+          writeFileSync(
+            settingsPath,
+            JSON.stringify({
+              security: { folderTrust: { enabled: false } },
+              ui: { useAlternateBuffer: true },
+            }),
+          );
+
+          const run = await rig.runInteractive({
+            args: ['--resume', 'latest'],
+            env: {
+              GEMINI_API_KEY: 'fake-perf-test-key',
+              GEMINI_TELEMETRY_ENABLED: 'true',
+              GEMINI_MEMORY_MONITOR_INTERVAL: '500',
+              GEMINI_EVENT_LOOP_MONITOR_ENABLED: 'true',
+              DEBUG: 'true',
+            },
+          });
+
+          await run.expectText('Type your message');
+
+          for (let i = 0; i < 5; i++) {
+            await run.sendKeys('\u001b[5~'); // PageUp
+          }
+
+          // Scroll to the very top
+          await run.sendKeys('\u001b[H'); // Home
+          // Verify top line of chat is visible.
+          await run.expectText('Authenticated with');
+
+          for (let i = 0; i < 5; i++) {
+            await run.sendKeys('\u001b[6~'); // PageDown
+          }
+
+          await rig.waitForTelemetryReady();
+          await run.kill();
+
+          const eventLoopMetric = rig.readMetric('event_loop.delay');
+          const cpuMetric = rig.readMetric('cpu.usage');
+
+          let p50Ms = 0;
+          let p95Ms = 0;
+          let maxMs = 0;
+          if (eventLoopMetric) {
+            const dataPoints = eventLoopMetric.dataPoints;
+            const p50Data = dataPoints.find(
+              (dp) => dp.attributes?.['percentile'] === 'p50',
+            );
+            const p95Data = dataPoints.find(
+              (dp) => dp.attributes?.['percentile'] === 'p95',
+            );
+            const maxData = dataPoints.find(
+              (dp) => dp.attributes?.['percentile'] === 'max',
+            );
+
+            if (p50Data?.value?.sum) p50Ms = p50Data.value.sum;
+            if (p95Data?.value?.sum) p95Ms = p95Data.value.sum;
+            if (maxData?.value?.sum) maxMs = maxData.value.sum;
+          }
+
+          let cpuTotalUs = 0;
+          if (cpuMetric) {
+            const dataPoints = cpuMetric.dataPoints;
+            for (const dp of dataPoints) {
+              if (dp.value?.sum && dp.value.sum > 0) {
+                cpuTotalUs += dp.value.sum;
+              }
+            }
+          }
+          const cpuUserUs = cpuTotalUs;
+          const cpuSystemUs = 0;
+
+          const snapshot: PerfSnapshot = {
+            timestamp: Date.now(),
+            label: 'scrolling',
+            wallClockMs: Math.round(p50Ms * 10) / 10,
+            cpuTotalUs,
+            cpuUserUs,
+            cpuSystemUs,
+            eventLoopDelayP50Ms: p50Ms,
+            eventLoopDelayP95Ms: p95Ms,
+            eventLoopDelayMaxMs: maxMs,
+          };
+
+          return snapshot;
+        },
+      );
+
+      if (UPDATE_BASELINES) {
+        harness.updateScenarioBaseline(result);
+      } else {
+        harness.assertWithinBaseline(result);
+      }
+    });
+  });
 });

perf-tests/perf.asian-language.responses

@@ -0,0 +1,2 @@
+{"method":"generateContent","response":{"candidates":[{"content":{"parts":[{"text":"0"}],"role":"model"},"finishReason":"STOP","index":0}]}}
+{"method":"generateContentStream","response":[{"candidates":[{"content":{"parts":[{"text":"你好！我是 Gemini CLI，你的 AI 编程助手"}],"role":"model"},"finishReason":"STOP","index":0}],"usageMetadata":{"promptTokenCount":20648,"candidatesTokenCount":12,"totalTokenCount":20769,"promptTokensDetails":[{"modality":"TEXT","tokenCount":5}]}}]}

perf-tests/perf.long-chat.responses

@@ -0,0 +1,4 @@
+{"method":"generateContent","response":{"candidates":[{"content":{"parts":[{"text":"{\"complexity_reasoning\":\"simple\",\"complexity_score\":1}"}],"role":"model"},"finishReason":"STOP","index":0}]}}
+{"method":"generateContentStream","response":[{"candidates":[{"content":{"parts":[{"text":"I am a large conversation model response."}],"role":"model"},"finishReason":"STOP","index":0}],"usageMetadata":{"candidatesTokenCount":10,"promptTokenCount":20,"totalTokenCount":30}}]}
+{"method":"generateContent","response":{"candidates":[{"content":{"parts":[{"text":"{\"originalSummary\":\"large chat summary\",\"events\":[]}"}],"role":"model"},"finishReason":"STOP","index":0}]}}
+{"method":"countTokens","response":{"totalTokens":100}}

scripts/copy_bundle_assets.js

@@ -108,4 +108,16 @@ if (!existsSync(bundleMcpSrc)) {
 cpSync(bundleMcpSrc, bundleMcpDest, { recursive: true, dereference: true });
 console.log('Copied bundled chrome-devtools-mcp to bundle/bundled/');
 
+// 7. Copy pre-built ripgrep vendor binaries
+const ripgrepVendorSrc = join(root, 'packages/core/vendor/ripgrep');
+const ripgrepVendorDest = join(bundleDir, 'vendor', 'ripgrep');
+if (existsSync(ripgrepVendorSrc)) {
+  mkdirSync(ripgrepVendorDest, { recursive: true });
+  cpSync(ripgrepVendorSrc, ripgrepVendorDest, {
+    recursive: true,
+    dereference: true,
+  });
+  console.log('Copied ripgrep vendor binaries to bundle/vendor/ripgrep/');
+}
+
 console.log('Assets copied to bundle/');

scripts/download-ripgrep-binaries.ts

@@ -0,0 +1,146 @@
+/**
+ * @license
+ * Copyright 2026 Google LLC
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+/**
+ * @fileoverview This script downloads pre-built ripgrep binaries for all supported
+ * architectures and platforms. These binaries are checked into the repository
+ * under packages/core/vendor/ripgrep.
+ *
+ * Maintainers should periodically run this script to upgrade the version
+ * of ripgrep being distributed.
+ *
+ * Usage: npx tsx scripts/download-ripgrep-binaries.ts
+ */
+
+import fs from 'node:fs';
+import fsPromises from 'node:fs/promises';
+import path from 'node:path';
+import { pipeline } from 'node:stream/promises';
+import { fileURLToPath } from 'node:url';
+import { createWriteStream } from 'node:fs';
+import { Readable } from 'node:stream';
+import type { ReadableStream } from 'node:stream/web';
+import { execFileSync } from 'node:child_process';
+
+const __dirname = path.dirname(fileURLToPath(import.meta.url));
+const CORE_VENDOR_DIR = path.join(__dirname, '../packages/core/vendor/ripgrep');
+const VERSION = 'v13.0.0-10';
+
+interface Target {
+  platform: string;
+  arch: string;
+  file: string;
+}
+
+const targets: Target[] = [
+  { platform: 'darwin', arch: 'arm64', file: 'aarch64-apple-darwin.tar.gz' },
+  { platform: 'darwin', arch: 'x64', file: 'x86_64-apple-darwin.tar.gz' },
+  {
+    platform: 'linux',
+    arch: 'arm64',
+    file: 'aarch64-unknown-linux-gnu.tar.gz',
+  },
+  { platform: 'linux', arch: 'x64', file: 'x86_64-unknown-linux-musl.tar.gz' },
+  { platform: 'win32', arch: 'x64', file: 'x86_64-pc-windows-msvc.zip' },
+];
+
+async function downloadBinary() {
+  await fsPromises.mkdir(CORE_VENDOR_DIR, { recursive: true });
+
+  for (const target of targets) {
+    const url = `https://github.com/microsoft/ripgrep-prebuilt/releases/download/${VERSION}/ripgrep-${VERSION}-${target.file}`;
+    const archivePath = path.join(CORE_VENDOR_DIR, target.file);
+    const binName = `rg-${target.platform}-${target.arch}${target.platform === 'win32' ? '.exe' : ''}`;
+    const finalBinPath = path.join(CORE_VENDOR_DIR, binName);
+
+    if (fs.existsSync(finalBinPath)) {
+      console.log(`[Cache] ${binName} already exists.`);
+      continue;
+    }
+
+    console.log(`[Download] ${url} -> ${archivePath}`);
+    const response = await fetch(url);
+    if (!response.ok) {
+      throw new Error(`Failed to fetch ${url}: ${response.statusText}`);
+    }
+
+    if (!response.body) {
+      throw new Error(`Response body is null for ${url}`);
+    }
+
+    const fileStream = createWriteStream(archivePath);
+
+    // Node 18+ global fetch response.body is a ReadableStream (web stream)
+    // pipeline(Readable.fromWeb(response.body), fileStream) works in Node 18+
+    await pipeline(
+      Readable.fromWeb(response.body as ReadableStream),
+      fileStream,
+    );
+
+    console.log(`[Extract] Extracting ${archivePath}...`);
+    // Extract using shell commands for simplicity
+    if (target.file.endsWith('.tar.gz')) {
+      execFileSync('tar', ['-xzf', archivePath, '-C', CORE_VENDOR_DIR]);
+      // Microsoft's ripgrep release extracts directly to `rg` inside the current directory sometimes
+      const sourceBin = path.join(CORE_VENDOR_DIR, 'rg');
+      if (fs.existsSync(sourceBin)) {
+        await fsPromises.rename(sourceBin, finalBinPath);
+      } else {
+        // Fallback for sub-directory if it happens
+        const extractedDirName = `ripgrep-${VERSION}-${target.file.replace('.tar.gz', '')}`;
+        const fallbackSourceBin = path.join(
+          CORE_VENDOR_DIR,
+          extractedDirName,
+          'rg',
+        );
+        if (fs.existsSync(fallbackSourceBin)) {
+          await fsPromises.rename(fallbackSourceBin, finalBinPath);
+          await fsPromises.rm(path.join(CORE_VENDOR_DIR, extractedDirName), {
+            recursive: true,
+            force: true,
+          });
+        } else {
+          throw new Error(
+            `Could not find extracted 'rg' binary for ${target.platform} ${target.arch}`,
+          );
+        }
+      }
+    } else if (target.file.endsWith('.zip')) {
+      execFileSync('unzip', ['-o', '-q', archivePath, '-d', CORE_VENDOR_DIR]);
+      const sourceBin = path.join(CORE_VENDOR_DIR, 'rg.exe');
+      if (fs.existsSync(sourceBin)) {
+        await fsPromises.rename(sourceBin, finalBinPath);
+      } else {
+        const extractedDirName = `ripgrep-${VERSION}-${target.file.replace('.zip', '')}`;
+        const fallbackSourceBin = path.join(
+          CORE_VENDOR_DIR,
+          extractedDirName,
+          'rg.exe',
+        );
+        if (fs.existsSync(fallbackSourceBin)) {
+          await fsPromises.rename(fallbackSourceBin, finalBinPath);
+          await fsPromises.rm(path.join(CORE_VENDOR_DIR, extractedDirName), {
+            recursive: true,
+            force: true,
+          });
+        } else {
+          throw new Error(
+            `Could not find extracted 'rg.exe' binary for ${target.platform} ${target.arch}`,
+          );
+        }
+      }
+    }
+
+    // Clean up archive
+    await fsPromises.unlink(archivePath);
+    console.log(`[Success] Saved to ${finalBinPath}`);
+  }
+}
+
+downloadBinary().catch((err) => {
+  console.error(err);
+  process.exit(1);
+});