6deaf3dd0f
fix(core,cli): stabilize agent harness, resolve build/lint errors, and fix flaky tests
2026-02-20 20:12:49 +00:00
102881c27f
feat(core): stabilize agent harness, fix result extraction and turn limits
2026-02-19 10:34:49 -06:00
456f8be568
wip
2026-02-18 15:45:46 -06:00
9d8351c5c9
wip: fork agent harness UI and fix engine delegation loop
2026-02-12 13:10:40 -06:00
3b9d5d6e8c
feat(core): add debug logging to legacy LocalAgentExecutor
2026-02-12 00:39:11 -06:00
303d001251
feat(core): add standardized lifecycle debug logging for agents
2026-02-12 00:31:56 -06:00
ed32dcb179
fix(core): truncate excessive debug output in MessageBus
2026-02-12 00:13:59 -06:00
39fe31d2d3
agent harnness
2026-02-11 22:31:43 -06:00
67819bf5ae
fix(core): ensure subagents have access to tools and skills in new harness
2026-02-11 21:53:12 -06:00
c989087ba5
refactor(core): unify ReAct loop in AgentHarness using Behavioral architecture
2026-02-11 21:28:57 -06:00
14e781c77c
refactor(core): unify ReAct loop in AgentHarness using Behavioral architecture
2026-02-11 21:19:11 -06:00
70105b687c
feat(core): implement recovery logic and time-based deadlines in AgentHarness
...
This adds DeadlineTimer support and a unified recovery loop to AgentHarness, bringing it to full parity with LocalAgentExecutor. (Skipped problematic lints for unsafe assertions on terminateReason enum)
2026-02-11 20:29:08 -06:00
3faa6f2056
fix(core): intercept complete_task calls in AgentHarness to prevent scheduler crash
...
Fixes a bug where the virtual complete_task tool was being passed to the tool scheduler, causing a 'Tool not found' error because it is not registered in the ToolRegistry. It is now handled internally by the harness.
2026-02-11 20:20:39 -06:00
85c31ad2da
feat(cli): add --experimental-agent-harness command line flag #18267
2026-02-11 17:43:37 -06:00
886417efc1
fix(core/cli): resolve build errors and integrate SubagentActivity event #18267
2026-02-11 17:33:38 -06:00
e650c10cf5
feat(core): implement unified AgentHarness and AgentFactory #18267
2026-02-11 17:24:20 -06:00
6c1773170e
More grep prompt tweaks ( #18846 )
2026-02-11 21:55:27 +00:00
4138667bae
feat(a2a): add value-resolver for auth credential resolution ( #18653 )
2026-02-11 21:23:28 +00:00
bfa791e13d
feat(core): update internal utility models to Gemini 3 ( #18773 )
2026-02-11 20:20:14 +00:00
e9a9474810
Revert unintended credentials exposure ( #18840 )
2026-02-11 20:06:28 +00:00
2a08456ed0
Update prompt and grep tool definition to limit context size ( #18780 )
2026-02-11 19:20:51 +00:00
84ce53aafa
feat(plan): allow skills to be enabled in plan mode ( #18817 )
...
Co-authored-by: Jerop Kipruto <jerop@google.com >
2026-02-11 17:59:03 +00:00
65d26e73a2
feat(plan): document and validate Plan Mode policy overrides ( #18825 )
2026-02-11 17:32:02 +00:00
0080589939
fix(cli): resolve double rendering in shpool and address vscode lint warnings ( #18704 )
2026-02-11 17:29:18 +00:00
34a47a51f4
fix(core): cache CLI version to ensure consistency during sessions ( #18793 )
2026-02-11 17:01:50 +00:00
f5dd1068f6
fix(core): complete MCP discovery when configured servers are skipped ( #18586 )
...
Co-authored-by: christine betts <chrstn@uw.edu >
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
2026-02-11 16:30:46 +00:00
5baad108d9
feat: multi-line text answers in ask-user tool ( #18741 )
2026-02-11 14:14:53 +00:00
3776c4d613
feat(masking): enable tool output masking by default ( #18564 )
2026-02-11 06:21:55 +00:00
0d034b8c18
Introduce limits for search results. ( #18767 )
2026-02-11 03:50:10 +00:00
49d55d972e
feat(core): formalize 5-phase sequential planning workflow ( #18759 )
2026-02-11 03:02:20 +00:00
6d3fff2ea4
fix(core): prevent race condition in policy persistence ( #18506 )
...
Co-authored-by: Allen Hutchison <adh@google.com >
2026-02-10 23:35:09 +00:00
cb4e1e684d
chore(core): update activate_skill prompt verbiage to be more direct ( #18605 )
2026-02-10 22:17:42 +00:00
b7a3243334
chore: cleanup unused and add unlisted dependencies in packages/core ( #18762 )
2026-02-10 22:07:06 +00:00
8257ec447a
chore(release): bump version to 0.30.0-nightly.20260210.a2174751d ( #18772 )
2026-02-10 17:13:00 -05:00
9590a092ae
Revert "fix(mcp): ensure MCP transport is closed to prevent memory leaks" ( #18771 )
2026-02-10 22:00:36 +00:00
a2174751de
feat(telemetry): Ensure experiment IDs are included in OpenTelemetry logs ( #18747 )
2026-02-10 21:01:35 +00:00
8b762111a8
Fix issue where Gemini CLI creates tests in a new file ( #18409 )
2026-02-10 20:53:29 +00:00
2eb1c92347
Fix issues with rip grep ( #18756 )
2026-02-10 20:48:56 +00:00
55571de066
feat: redact disabled tools from system prompt ( #13597 ) ( #18613 )
2026-02-10 19:00:36 +00:00
740f0e4c3d
fix: allow ask_user tool in yolo mode ( #18541 )
2026-02-10 18:56:51 +00:00
b37e67451a
ci: shard windows tests and fix event listener leaks ( #18670 )
2026-02-10 18:46:42 +00:00
262138cad5
test: add model-specific snapshots for coreTools ( #18707 )
...
Co-authored-by: matt korwel <matt.korwel@gmail.com >
2026-02-10 18:41:01 +00:00
41bbe6ca0a
fix(core): standardize tool formatting in system prompts ( #18615 )
2026-02-10 15:30:08 +00:00
37f128a109
feat(routing): restrict numerical routing to Gemini 3 family ( #18478 )
2026-02-10 14:25:21 +00:00
2ae5e1ae20
feat(core): optimize sub-agents system prompt intro ( #18608 )
2026-02-10 08:25:42 +00:00
92a5f725a1
refactor(core): refine Security & System Integrity section in system prompt ( #18601 )
2026-02-10 04:32:36 +00:00
6dae3a5402
Feature/quota visibility 16795 ( #18203 )
2026-02-10 02:53:10 +00:00
89d4556c45
feat(core): Render memory hierarchically in context. ( #18350 )
2026-02-10 02:01:59 +00:00
5d0570b113
refactor(core): centralize core tool definitions and support model-specific schemas ( #18662 )
2026-02-10 01:29:52 +00:00
cc2798018b
feat: handle multiple dynamic context filenames in system prompt ( #18598 )
2026-02-10 00:37:08 +00:00
mkorwel