Merge branch 'pr4-cli' into mb/agent-interactive

Merge branch 'main' into pr4-cli
refactor(cli): remove config dependency from useAgentStream
2026-07-23 08:10:57 -07:00 · 2026-03-30 12:59:33 -07:00 · 2026-03-30 12:39:30 -07:00 · 2026-03-30 12:35:31 -07:00 · 2026-03-30 09:48:31 -07:00 · 2026-03-30 14:31:20 +00:00
274 changed files with 12142 additions and 4622 deletions
@@ -0,0 +1,89 @@
+# --- STAGE 1: Base Runtime ---
+FROM docker.io/library/node:20-slim AS base
+
+RUN apt-get update && apt-get install -y --no-install-recommends \
+  python3 \
+  python3-pip \
+  python3-venv \
+  curl \
+  dnsutils \
+  less \
+  jq \
+  ca-certificates \
+  git \
+  && apt-get clean \
+  && rm -rf /var/lib/apt/lists/*
+
+# --- STAGE 2: Builder (Compile Main) ---
+FROM base AS builder
+WORKDIR /build
+COPY . .
+RUN npm ci --ignore-scripts
+RUN npm run bundle
+# Run the official release preparation script to move the bundle and assets into packages/cli
+RUN node scripts/prepare-npm-release.js
+
+# --- STAGE 3: Development Environment ---
+FROM base AS development
+
+WORKDIR /home/node/dev/main
+
+# Set up npm global package folder
+RUN mkdir -p /usr/local/share/npm-global \
+  && chown -R node:node /usr/local/share/npm-global
+ENV NPM_CONFIG_PREFIX=/usr/local/share/npm-global
+ENV PATH=$PATH:/usr/local/share/npm-global/bin
+
+# Copy package.json to extract versions for global tools
+COPY package.json /tmp/package.json
+
+# Install Build Tools, Global Dev Tools (pinned), and Linters
+ARG ACTIONLINT_VER=1.7.7
+ARG SHELLCHECK_VER=0.11.0
+ARG YAMLLINT_VER=1.35.1
+
+RUN apt-get update && apt-get install -y --no-install-recommends \
+  make \
+  g++ \
+  gh \
+  git \
+  unzip \
+  rsync \
+  ripgrep \
+  procps \
+  psmisc \
+  lsof \
+  socat \
+  tmux \
+  docker.io \
+  build-essential \
+  libsecret-1-dev \
+  libkrb5-dev \
+  file \
+  && curl -sSLo /tmp/actionlint.tar.gz https://github.com/rhysd/actionlint/releases/download/v${ACTIONLINT_VER}/actionlint_${ACTIONLINT_VER}_linux_amd64.tar.gz \
+  && tar -xzf /tmp/actionlint.tar.gz -C /usr/local/bin actionlint \
+  && curl -sSLo /tmp/shellcheck.tar.xz https://github.com/koalaman/shellcheck/releases/download/v${SHELLCHECK_VER}/shellcheck-v${SHELLCHECK_VER}.linux.x86_64.tar.xz \
+  && tar -xf /tmp/shellcheck.tar.xz -C /usr/local/bin --strip-components=1 shellcheck-v${SHELLCHECK_VER}/shellcheck \
+  && pip3 install --break-system-packages yamllint==${YAMLLINT_VER} \
+  && export TSX_VER=$(node -p "require('/tmp/package.json').devDependencies.tsx") \
+  && export VITEST_VER=$(node -p "require('/tmp/package.json').devDependencies.vitest") \
+  && export PRETTIER_VER=$(node -p "require('/tmp/package.json').devDependencies.prettier") \
+  && export ESLINT_VER=$(node -p "require('/tmp/package.json').devDependencies.eslint") \
+  && export CROSS_ENV_VER=$(node -p "require('/tmp/package.json').devDependencies['cross-env']") \
+  && npm install -g tsx@$TSX_VER vitest@$VITEST_VER prettier@$PRETTIER_VER eslint@$ESLINT_VER cross-env@$CROSS_ENV_VER typescript@5.3.3 \
+  && npm install -g @google/gemini-cli@nightly && mv /usr/local/share/npm-global/bin/gemini /usr/local/share/npm-global/bin/g-nightly \
+  && npm install -g @google/gemini-cli@preview && mv /usr/local/share/npm-global/bin/gemini /usr/local/share/npm-global/bin/g-preview \
+  && npm install -g @google/gemini-cli@latest  && mv /usr/local/share/npm-global/bin/gemini /usr/local/share/npm-global/bin/g-stable \
+  && apt-get purge -y build-essential libsecret-1-dev libkrb5-dev \
+  && apt-get autoremove -y \
+  && apt-get clean \
+  && rm -rf /var/lib/apt/lists/* /tmp/* /root/.npm
+
+# Copy the bundled CLI package to a permanent location and install it
+# We MUST not delete this source folder as 'npm install -g <folder>' 
+# often symlinks to it for local folder installs.
+COPY --from=builder /build/packages/cli /usr/local/lib/gemini-cli
+RUN npm install -g /usr/local/lib/gemini-cli
+
+USER node
+CMD ["/bin/bash"]
@@ -0,0 +1,10 @@
+node_modules
+.git
+.gemini/workspaces
+dist
+!packages/*/dist/*.tgz
+bundle
+out
+*.log
+.env
+.DS_Store
@@ -0,0 +1,58 @@
+substitutions:
+  _IMAGE_NAME: 'development'
+  _ARTIFACT_REGISTRY_REPO: 'us-docker.pkg.dev/gemini-code-dev/gemini-cli'
+
+steps:
+  # Step 1: Install root dependencies
+  - name: 'us-west1-docker.pkg.dev/gemini-code-dev/gemini-code-containers/gemini-code-builder'
+    id: 'Install Dependencies'
+    entrypoint: 'npm'
+    args: ['install']
+
+  # Step 2: Authenticate for Docker
+  - name: 'us-west1-docker.pkg.dev/gemini-code-dev/gemini-code-containers/gemini-code-builder'
+    id: 'Authenticate docker'
+    entrypoint: 'npm'
+    args: ['run', 'auth']
+
+  # Step 3: Build workspace packages
+  - name: 'us-west1-docker.pkg.dev/gemini-code-dev/gemini-code-containers/gemini-code-builder'
+    id: 'Build packages'
+    entrypoint: 'npm'
+    args: ['run', 'build:packages']
+
+  # Step 4: Build Development Image
+  - name: 'us-west1-docker.pkg.dev/gemini-code-dev/gemini-code-containers/gemini-code-builder'
+    id: 'Build Development Image'
+    entrypoint: 'bash'
+    env:
+      - 'RAW_BRANCH_VALUE=${BRANCH_NAME}'
+    args:
+      - '-c'
+      - |-
+        IMAGE_BASE="${_ARTIFACT_REGISTRY_REPO}/${_IMAGE_NAME}"
+
+        # Determine the primary tag (branch name or 'latest' for main)
+        # Use $$ for shell variables to avoid Cloud Build attempting premature substitution
+        RAW_BRANCH="$$RAW_BRANCH_VALUE"
+        if [ "$${RAW_BRANCH}" == "main" ]; then
+          TAG_PRIMARY="latest"
+        else
+          TAG_PRIMARY=$$(echo "$${RAW_BRANCH}" | sed 's/[^a-zA-Z0-9]/-/g' | tr '[:upper:]' '[:lower:]')
+        fi
+
+        # Use SHORT_SHA if available (Cloud Build) or fallback to latest-dev
+        TAG_SHA="$${SHORT_SHA:-latest-dev}"
+
+        echo "📦 Building Development Image for: $${RAW_BRANCH} -> $${TAG_PRIMARY} ($${TAG_SHA})"
+
+        docker build -f .gcp/Dockerfile.development \
+          -t "$${IMAGE_BASE}:$${TAG_SHA}" \
+          -t "$${IMAGE_BASE}:$${TAG_PRIMARY}" .
+
+        docker push "$${IMAGE_BASE}:$${TAG_SHA}"
+        docker push "$${IMAGE_BASE}:$${TAG_PRIMARY}"
+
+options:
+  defaultLogsBucketBehavior: 'REGIONAL_USER_OWNED_BUCKET'
+  dynamicSubstitutions: true
@@ -1,190 +1,194 @@
 #!/bin/bash

 notify() {
-  local title="$1"
-  local message="$2"
-  local pr="$3"
+  local title="${1}"
+  local message="${2}"
+  local pr="${3}"
  # Terminal escape sequence
-  printf "\e]9;%s | PR #%s | %s\a" "$title" "$pr" "$message"
+  printf "\e]9;%s | PR #%s | %s\a" "${title}" "${pr}" "${message}"
  # Native macOS notification
-  if [[ "$(uname)" == "Darwin" ]]; then
-    osascript -e "display notification \"$message\" with title \"$title\" subtitle \"PR #$pr\""
+  os_type="$(uname || true)"
+  if [[ "${os_type}" == "Darwin" ]]; then
+    osascript -e "display notification \"${message}\" with title \"${title}\" subtitle \"PR #${pr}\""
  fi
 }

-pr_number=$1
-if [[ -z "$pr_number" ]]; then
+pr_number="${1}"
+if [[ -z "${pr_number}" ]]; then
  echo "Usage: async-review <pr_number>"
  exit 1
 fi

-base_dir=$(git rev-parse --show-toplevel 2>/dev/null)
-if [[ -z "$base_dir" ]]; then
+base_dir="$(git rev-parse --show-toplevel 2>/dev/null || true)"
+if [[ -z "${base_dir}" ]]; then
  echo "❌ Must be run from within a git repository."
  exit 1
 fi

 # Use the repository's local .gemini/tmp directory for ephemeral worktrees and logs
-pr_dir="$base_dir/.gemini/tmp/async-reviews/pr-$pr_number"
-target_dir="$pr_dir/worktree"
-log_dir="$pr_dir/logs"
+pr_dir="${base_dir}/.gemini/tmp/async-reviews/pr-${pr_number}"
+target_dir="${pr_dir}/worktree"
+log_dir="${pr_dir}/logs"

-cd "$base_dir" || exit 1
+cd "${base_dir}" || exit 1

-mkdir -p "$log_dir"
-rm -f "$log_dir/setup.exit" "$log_dir/final-assessment.exit" "$log_dir/final-assessment.md"
+mkdir -p "${log_dir}"
+rm -f "${log_dir}/setup.exit" "${log_dir}/final-assessment.exit" "${log_dir}/final-assessment.md"

-echo "🧹 Cleaning up previous worktree if it exists..." | tee -a "$log_dir/setup.log"
-git worktree remove -f "$target_dir" >> "$log_dir/setup.log" 2>&1 || true
-git branch -D "gemini-async-pr-$pr_number" >> "$log_dir/setup.log" 2>&1 || true
-git worktree prune >> "$log_dir/setup.log" 2>&1 || true
+echo "🧹 Cleaning up previous worktree if it exists..." | tee -a "${log_dir}/setup.log"
+git worktree remove -f "${target_dir}" >> "${log_dir}/setup.log" 2>&1 || true
+git branch -D "gemini-async-pr-${pr_number}" >> "${log_dir}/setup.log" 2>&1 || true
+git worktree prune >> "${log_dir}/setup.log" 2>&1 || true

-echo "📡 Fetching PR #$pr_number..." | tee -a "$log_dir/setup.log"
-if ! git fetch origin -f "pull/$pr_number/head:gemini-async-pr-$pr_number" >> "$log_dir/setup.log" 2>&1; then
-  echo 1 > "$log_dir/setup.exit"
-  echo "❌ Fetch failed. Check $log_dir/setup.log"
-  notify "Async Review Failed" "Fetch failed." "$pr_number"
+echo "📡 Fetching PR #${pr_number}..." | tee -a "${log_dir}/setup.log"
+if ! git fetch origin -f "pull/${pr_number}/head:gemini-async-pr-${pr_number}" >> "${log_dir}/setup.log" 2>&1; then
+  echo 1 > "${log_dir}/setup.exit"
+  echo "❌ Fetch failed. Check ${log_dir}/setup.log"
+  notify "Async Review Failed" "Fetch failed." "${pr_number}"
  exit 1
 fi

-if [[ ! -d "$target_dir" ]]; then
-  echo "🧹 Pruning missing worktrees..." | tee -a "$log_dir/setup.log"
-  git worktree prune >> "$log_dir/setup.log" 2>&1
-  echo "🌿 Creating worktree in $target_dir..." | tee -a "$log_dir/setup.log"
-  if ! git worktree add "$target_dir" "gemini-async-pr-$pr_number" >> "$log_dir/setup.log" 2>&1; then
-    echo 1 > "$log_dir/setup.exit"
-    echo "❌ Worktree creation failed. Check $log_dir/setup.log"
-    notify "Async Review Failed" "Worktree creation failed." "$pr_number"
+if [[ ! -d "${target_dir}" ]]; then
+  echo "🧹 Pruning missing worktrees..." | tee -a "${log_dir}/setup.log"
+  git worktree prune >> "${log_dir}/setup.log" 2>&1
+  echo "🌿 Creating worktree in ${target_dir}..." | tee -a "${log_dir}/setup.log"
+  if ! git worktree add "${target_dir}" "gemini-async-pr-${pr_number}" >> "${log_dir}/setup.log" 2>&1; then
+    echo 1 > "${log_dir}/setup.exit"
+    echo "❌ Worktree creation failed. Check ${log_dir}/setup.log"
+    notify "Async Review Failed" "Worktree creation failed." "${pr_number}"
    exit 1
  fi
 else
-  echo "🌿 Worktree already exists." | tee -a "$log_dir/setup.log"
+  echo "🌿 Worktree already exists." | tee -a "${log_dir}/setup.log"
 fi
-echo 0 > "$log_dir/setup.exit"
+echo 0 > "${log_dir}/setup.exit"

-cd "$target_dir" || exit 1
+cd "${target_dir}" || exit 1

-echo "🚀 Launching background tasks. Logs saving to: $log_dir"
+echo "🚀 Launching background tasks. Logs saving to: ${log_dir}"

 echo "  ↳ [1/5] Grabbing PR diff..."
-rm -f "$log_dir/pr-diff.exit"
-{ gh pr diff "$pr_number" > "$log_dir/pr-diff.diff" 2>&1; echo $? > "$log_dir/pr-diff.exit"; } &
+rm -f "${log_dir}/pr-diff.exit"
+{ gh pr diff "${pr_number}" > "${log_dir}/pr-diff.diff" 2>&1; echo $? > "${log_dir}/pr-diff.exit"; } &

 echo "  ↳ [2/5] Starting build and lint..."
-rm -f "$log_dir/build-and-lint.exit"
-{ { npm run clean && npm ci && npm run format && npm run build && npm run lint:ci && npm run typecheck; } > "$log_dir/build-and-lint.log" 2>&1; echo $? > "$log_dir/build-and-lint.exit"; } &
+rm -f "${log_dir}/build-and-lint.exit"
+{ { npm run clean && npm ci && npm run format && npm run build && npm run lint:ci && npm run typecheck; } > "${log_dir}/build-and-lint.log" 2>&1; echo $? > "${log_dir}/build-and-lint.exit"; } &

 # Dynamically resolve gemini binary (fallback to your nightly path)
-GEMINI_CMD=$(which gemini || echo "$HOME/.gcli/nightly/node_modules/.bin/gemini")
+GEMINI_CMD="$(command -v gemini || echo "${HOME}/.gcli/nightly/node_modules/.bin/gemini")"
+# shellcheck disable=SC2312
 POLICY_PATH="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)/policy.toml"

 echo "  ↳ [3/5] Starting Gemini code review..."
-rm -f "$log_dir/review.exit"
-{ "$GEMINI_CMD" --policy "$POLICY_PATH" -p "/review-frontend $pr_number" > "$log_dir/review.md" 2>&1; echo $? > "$log_dir/review.exit"; } &
+rm -f "${log_dir}/review.exit"
+{ "${GEMINI_CMD}" --policy "${POLICY_PATH}" -p "/review-frontend ${pr_number}" > "${log_dir}/review.md" 2>&1; echo $? > "${log_dir}/review.exit"; } &

 echo "  ↳ [4/5] Starting automated tests (waiting for build and lint)..."
-rm -f "$log_dir/npm-test.exit"
+rm -f "${log_dir}/npm-test.exit"
 { 
-  while [ ! -f "$log_dir/build-and-lint.exit" ]; do sleep 1; done
-  if [ "$(cat "$log_dir/build-and-lint.exit")" == "0" ]; then
-    gh pr checks "$pr_number" > "$log_dir/ci-checks.log" 2>&1
+  while [[ ! -f "${log_dir}/build-and-lint.exit" ]]; do sleep 1; done
+  read -r build_exit < "${log_dir}/build-and-lint.exit" || build_exit=""
+  if [[ "${build_exit}" == "0" ]]; then
+    gh pr checks "${pr_number}" > "${log_dir}/ci-checks.log" 2>&1
    ci_status=$?
    
-    if [ "$ci_status" -eq 0 ]; then
-      echo "CI checks passed. Skipping local npm tests." > "$log_dir/npm-test.log"
-      echo 0 > "$log_dir/npm-test.exit"
-    elif [ "$ci_status" -eq 8 ]; then
-      echo "CI checks are still pending. Skipping local npm tests to avoid duplicate work. Please check GitHub for final results." > "$log_dir/npm-test.log"
-      echo 0 > "$log_dir/npm-test.exit"
+    if [[ "${ci_status}" -eq 0 ]]; then
+      echo "CI checks passed. Skipping local npm tests." > "${log_dir}/npm-test.log"
+      echo 0 > "${log_dir}/npm-test.exit"
+    elif [[ "${ci_status}" -eq 8 ]]; then
+      echo "CI checks are still pending. Skipping local npm tests to avoid duplicate work. Please check GitHub for final results." > "${log_dir}/npm-test.log"
+      echo 0 > "${log_dir}/npm-test.exit"
    else
-      echo "CI checks failed. Failing checks:" > "$log_dir/npm-test.log"
-      gh pr checks "$pr_number" --json name,bucket -q '.[] | select(.bucket=="fail") | .name' >> "$log_dir/npm-test.log" 2>&1
+      echo "CI checks failed. Failing checks:" > "${log_dir}/npm-test.log"
+      gh pr checks "${pr_number}" --json name,bucket -q '.[] | select(.bucket=="fail") | .name' >> "${log_dir}/npm-test.log" 2>&1
      
-      echo "Attempting to extract failing test files from CI logs..." >> "$log_dir/npm-test.log"
-      pr_branch=$(gh pr view "$pr_number" --json headRefName -q '.headRefName' 2>/dev/null)
-      run_id=$(gh run list --branch "$pr_branch" --workflow ci.yml --json databaseId -q '.[0].databaseId' 2>/dev/null)
+      echo "Attempting to extract failing test files from CI logs..." >> "${log_dir}/npm-test.log"
+      pr_branch="$(gh pr view "${pr_number}" --json headRefName -q '.headRefName' 2>/dev/null || true)"
+      run_id="$(gh run list --branch "${pr_branch}" --workflow ci.yml --json databaseId -q '.[0].databaseId' 2>/dev/null || true)"
      
      failed_files=""
-      if [[ -n "$run_id" ]]; then
-        failed_files=$(gh run view "$run_id" --log-failed 2>/dev/null | grep -o -E '(packages/[a-zA-Z0-9_-]+|integration-tests|evals)/[a-zA-Z0-9_/-]+\.test\.ts(x)?' | sort | uniq)
+      if [[ -n "${run_id}" ]]; then
+        failed_files="$(gh run view "${run_id}" --log-failed 2>/dev/null | grep -o -E '(packages/[a-zA-Z0-9_-]+|integration-tests|evals)/[a-zA-Z0-9_/-]+\.test\.ts(x)?' | sort | uniq || true)"
      fi
      
-      if [[ -n "$failed_files" ]]; then
-        echo "Found failing test files from CI:" >> "$log_dir/npm-test.log"
-        for f in $failed_files; do echo "  - $f" >> "$log_dir/npm-test.log"; done
-        echo "Running ONLY failing tests locally..." >> "$log_dir/npm-test.log"
+      if [[ -n "${failed_files}" ]]; then
+        echo "Found failing test files from CI:" >> "${log_dir}/npm-test.log"
+        for f in ${failed_files}; do echo "  - ${f}" >> "${log_dir}/npm-test.log"; done
+        echo "Running ONLY failing tests locally..." >> "${log_dir}/npm-test.log"
        
        exit_code=0
-        for file in $failed_files; do
-           if [[ "$file" == packages/* ]]; then
-             ws_dir=$(echo "$file" | cut -d'/' -f1,2)
+        for file in ${failed_files}; do
+           if [[ "${file}" == packages/* ]]; then
+             ws_dir="$(echo "${file}" | cut -d'/' -f1,2)"
           else
-             ws_dir=$(echo "$file" | cut -d'/' -f1)
+             ws_dir="$(echo "${file}" | cut -d'/' -f1)"
           fi
-           rel_file=${file#$ws_dir/}
+           rel_file="${file#"${ws_dir}"/}"
           
-           echo "--- Running $rel_file in workspace $ws_dir ---" >> "$log_dir/npm-test.log"
-           if ! npm run test:ci -w "$ws_dir" -- "$rel_file" >> "$log_dir/npm-test.log" 2>&1; then
+           echo "--- Running ${rel_file} in workspace ${ws_dir} ---" >> "${log_dir}/npm-test.log"
+           if ! npm run test:ci -w "${ws_dir}" -- "${rel_file}" >> "${log_dir}/npm-test.log" 2>&1; then
             exit_code=1
           fi
        done
-        echo $exit_code > "$log_dir/npm-test.exit"
+        echo "${exit_code}" > "${log_dir}/npm-test.exit"
      else
-        echo "Could not extract specific failing files. Skipping full local test suite as it takes too long. Please check CI logs manually." >> "$log_dir/npm-test.log"
-        echo 1 > "$log_dir/npm-test.exit"
+        echo "Could not extract specific failing files. Skipping full local test suite as it takes too long. Please check CI logs manually." >> "${log_dir}/npm-test.log"
+        echo 1 > "${log_dir}/npm-test.exit"
      fi
    fi
  else
-    echo "Skipped due to build-and-lint failure" > "$log_dir/npm-test.log"
-    echo 1 > "$log_dir/npm-test.exit"
+    echo "Skipped due to build-and-lint failure" > "${log_dir}/npm-test.log"
+    echo 1 > "${log_dir}/npm-test.exit"
  fi
 } &

 echo "  ↳ [5/5] Starting Gemini test execution (waiting for build and lint)..."
-rm -f "$log_dir/test-execution.exit"
+rm -f "${log_dir}/test-execution.exit"
 { 
-  while [ ! -f "$log_dir/build-and-lint.exit" ]; do sleep 1; done
-  if [ "$(cat "$log_dir/build-and-lint.exit")" == "0" ]; then
-    "$GEMINI_CMD" --policy "$POLICY_PATH" -p "Analyze the diff for PR $pr_number using 'gh pr diff $pr_number'. Instead of running the project's automated test suite (like 'npm test'), physically exercise the newly changed code in the terminal (e.g., by writing a temporary script to call the new functions, or testing the CLI command directly). Verify the feature's behavior works as expected. IMPORTANT: Do NOT modify any source code to fix errors. Just exercise the code and log the results, reporting any failures clearly. Do not ask for user confirmation." > "$log_dir/test-execution.log" 2>&1; echo $? > "$log_dir/test-execution.exit"
+  while [[ ! -f "${log_dir}/build-and-lint.exit" ]]; do sleep 1; done
+  read -r build_exit < "${log_dir}/build-and-lint.exit" || build_exit=""
+  if [[ "${build_exit}" == "0" ]]; then
+    "${GEMINI_CMD}" --policy "${POLICY_PATH}" -p "Analyze the diff for PR ${pr_number} using 'gh pr diff ${pr_number}'. Instead of running the project's automated test suite (like 'npm test'), physically exercise the newly changed code in the terminal (e.g., by writing a temporary script to call the new functions, or testing the CLI command directly). Verify the feature's behavior works as expected. IMPORTANT: Do NOT modify any source code to fix errors. Just exercise the code and log the results, reporting any failures clearly. Do not ask for user confirmation." > "${log_dir}/test-execution.log" 2>&1; echo $? > "${log_dir}/test-execution.exit"
  else
-    echo "Skipped due to build-and-lint failure" > "$log_dir/test-execution.log"
-    echo 1 > "$log_dir/test-execution.exit"
+    echo "Skipped due to build-and-lint failure" > "${log_dir}/test-execution.log"
+    echo 1 > "${log_dir}/test-execution.exit"
  fi
 } &

 echo "✅ All tasks dispatched!"
-echo "You can monitor progress with: tail -f $log_dir/*.log"
-echo "Read your review later at: $log_dir/review.md"
+echo "You can monitor progress with: tail -f ${log_dir}/*.log"
+echo "Read your review later at: ${log_dir}/review.md"

 # Polling loop to wait for all background tasks to finish
 tasks=("pr-diff" "build-and-lint" "review" "npm-test" "test-execution")
 log_files=("pr-diff.diff" "build-and-lint.log" "review.md" "npm-test.log" "test-execution.log")

 declare -A task_done
-for t in "${tasks[@]}"; do task_done[$t]=0; done
+for t in "${tasks[@]}"; do task_done[${t}]=0; done

 all_done=0
-while [[ $all_done -eq 0 ]]; do
+while [[ "${all_done}" -eq 0 ]]; do
  clear
  echo "=================================================="
-  echo "🚀 Async PR Review Status for PR #$pr_number"
+  echo "🚀 Async PR Review Status for PR #${pr_number}"
  echo "=================================================="
  echo ""
  
  all_done=1
  for i in "${!tasks[@]}"; do
-    t="${tasks[$i]}"
+    t="${tasks[${i}]}"
    
-    if [[ -f "$log_dir/$t.exit" ]]; then
-      exit_code=$(cat "$log_dir/$t.exit")
-      if [[ "$exit_code" == "0" ]]; then
-        echo "  ✅ $t: SUCCESS"
+    if [[ -f "${log_dir}/${t}.exit" ]]; then
+      read -r task_exit < "${log_dir}/${t}.exit" || task_exit=""
+      if [[ "${task_exit}" == "0" ]]; then
+        echo "  ✅ ${t}: SUCCESS"
      else
-        echo "  ❌ $t: FAILED (exit code $exit_code)"
+        echo "  ❌ ${t}: FAILED (exit code ${task_exit})"
      fi
-      task_done[$t]=1
+      task_done[${t}]=1
    else
-      echo "  ⏳ $t: RUNNING"
+      echo "  ⏳ ${t}: RUNNING"
      all_done=0
    fi
  done
@@ -195,47 +199,47 @@ while [[ $all_done -eq 0 ]]; do
  echo "=================================================="
  
  for i in "${!tasks[@]}"; do
-    t="${tasks[$i]}"
-    log_file="${log_files[$i]}"
+    t="${tasks[${i}]}"
+    log_file="${log_files[${i}]}"
    
-    if [[ ${task_done[$t]} -eq 0 ]]; then
-      if [[ -f "$log_dir/$log_file" ]]; then
+    if [[ "${task_done[${t}]}" -eq 0 ]]; then
+      if [[ -f "${log_dir}/${log_file}" ]]; then
        echo ""
-        echo "--- $t ---"
-        tail -n 5 "$log_dir/$log_file"
+        echo "--- ${t} ---"
+        tail -n 5 "${log_dir}/${log_file}"
      fi
    fi
  done
  
-  if [[ $all_done -eq 0 ]]; then
+  if [[ "${all_done}" -eq 0 ]]; then
    sleep 3
  fi
 done

 clear
 echo "=================================================="
-echo "🚀 Async PR Review Status for PR #$pr_number"
+echo "🚀 Async PR Review Status for PR #${pr_number}"
 echo "=================================================="
 echo ""
 for t in "${tasks[@]}"; do
-  exit_code=$(cat "$log_dir/$t.exit")
-  if [[ "$exit_code" == "0" ]]; then
-    echo "  ✅ $t: SUCCESS"
+  read -r task_exit < "${log_dir}/${t}.exit" || task_exit=""
+  if [[ "${task_exit}" == "0" ]]; then
+    echo "  ✅ ${t}: SUCCESS"
  else
-    echo "  ❌ $t: FAILED (exit code $exit_code)"
+    echo "  ❌ ${t}: FAILED (exit code ${task_exit})"
  fi
 done
 echo ""

 echo "⏳ Tasks complete! Synthesizing final assessment..."
-if ! "$GEMINI_CMD" --policy "$POLICY_PATH" -p "Read the review at $log_dir/review.md, the automated test logs at $log_dir/npm-test.log, and the manual test execution logs at $log_dir/test-execution.log. Summarize the results, state whether the build and tests passed based on $log_dir/build-and-lint.exit and $log_dir/npm-test.exit, and give a final recommendation for PR $pr_number." > "$log_dir/final-assessment.md" 2>&1; then
-  echo $? > "$log_dir/final-assessment.exit"
+if ! "${GEMINI_CMD}" --policy "${POLICY_PATH}" -p "Read the review at ${log_dir}/review.md, the automated test logs at ${log_dir}/npm-test.log, and the manual test execution logs at ${log_dir}/test-execution.log. Summarize the results, state whether the build and tests passed based on ${log_dir}/build-and-lint.exit and ${log_dir}/npm-test.exit, and give a final recommendation for PR ${pr_number}." > "${log_dir}/final-assessment.md" 2>&1; then
+  echo $? > "${log_dir}/final-assessment.exit"
  echo "❌ Final assessment synthesis failed!"
-  echo "Check $log_dir/final-assessment.md for details."
-  notify "Async Review Failed" "Final assessment synthesis failed." "$pr_number"
+  echo "Check ${log_dir}/final-assessment.md for details."
+  notify "Async Review Failed" "Final assessment synthesis failed." "${pr_number}"
  exit 1
 fi

-echo 0 > "$log_dir/final-assessment.exit"
-echo "✅ Final assessment complete! Check $log_dir/final-assessment.md"
-notify "Async Review Complete" "Review and test execution finished successfully." "$pr_number"
+echo 0 > "${log_dir}/final-assessment.exit"
+echo "✅ Final assessment complete! Check ${log_dir}/final-assessment.md"
+notify "Async Review Complete" "Review and test execution finished successfully." "${pr_number}"
@@ -1,22 +1,22 @@
 #!/bin/bash
-pr_number=$1
+pr_number="${1}"

-if [[ -z "$pr_number" ]]; then
+if [[ -z "${pr_number}" ]]; then
  echo "Usage: check-async-review <pr_number>"
  exit 1
 fi

-base_dir=$(git rev-parse --show-toplevel 2>/dev/null)
-if [[ -z "$base_dir" ]]; then
+base_dir="$(git rev-parse --show-toplevel 2>/dev/null || true)"
+if [[ -z "${base_dir}" ]]; then
  echo "❌ Must be run from within a git repository."
  exit 1
 fi

-log_dir="$base_dir/.gemini/tmp/async-reviews/pr-$pr_number/logs"
+log_dir="${base_dir}/.gemini/tmp/async-reviews/pr-${pr_number}/logs"

-if [[ ! -d "$log_dir" ]]; then
+if [[ ! -d "${log_dir}" ]]; then
  echo "STATUS: NOT_FOUND"
-  echo "❌ No logs found for PR #$pr_number in $log_dir"
+  echo "❌ No logs found for PR #${pr_number} in ${log_dir}"
  exit 0
 fi

@@ -34,32 +34,32 @@ all_done=true
 echo "STATUS: CHECKING"

 for task_info in "${tasks[@]}"; do
-  IFS="|" read -r task_name log_file <<< "$task_info"
+  IFS="|" read -r task_name log_file <<< "${task_info}"
  
-  file_path="$log_dir/$log_file"
-  exit_file="$log_dir/$task_name.exit"
+  file_path="${log_dir}/${log_file}"
+  exit_file="${log_dir}/${task_name}.exit"

-  if [[ -f "$exit_file" ]]; then
-    exit_code=$(cat "$exit_file")
-    if [[ "$exit_code" == "0" ]]; then
-      echo "✅ $task_name: SUCCESS"
+  if [[ -f "${exit_file}" ]]; then
+    read -r exit_code < "${exit_file}" || exit_code=""
+    if [[ "${exit_code}" == "0" ]]; then
+      echo "✅ ${task_name}: SUCCESS"
    else
-      echo "❌ $task_name: FAILED (exit code $exit_code)"
-      echo "   Last lines of $file_path:"
-      tail -n 3 "$file_path" | sed 's/^/      /'
+      echo "❌ ${task_name}: FAILED (exit code ${exit_code})"
+      echo "   Last lines of ${file_path}:"
+      tail -n 3 "${file_path}" | sed 's/^/      /' || true
    fi
-  elif [[ -f "$file_path" ]]; then
-    echo "⏳ $task_name: RUNNING"
+  elif [[ -f "${file_path}" ]]; then
+    echo "⏳ ${task_name}: RUNNING"
    all_done=false
  else
-    echo "➖ $task_name: NOT STARTED"
+    echo "➖ ${task_name}: NOT STARTED"
    all_done=false
  fi
 done

-if $all_done; then
+if [[ "${all_done}" == "true" ]]; then
  echo "STATUS: COMPLETE"
-  echo "LOG_DIR: $log_dir"
+  echo "LOG_DIR: ${log_dir}"
 else
  echo "STATUS: IN_PROGRESS"
-fi
+fi
@@ -65,8 +65,6 @@ accessible.
 - **UI and code:** Use **bold** for UI elements and `code font` for filenames,
  snippets, commands, and API elements. Focus on the task when discussing
  interaction.
- **Links:** Use descriptive anchor text; avoid "click here." Ensure the link
-  makes sense out of context.
 - **Accessibility:** Use semantic HTML elements correctly (headings, lists, 
  tables).
 - **Media:** Use lowercase hyphenated filenames. Provide descriptive alt text
@@ -100,6 +98,18 @@ accessible.
 > This is an example of a multi-line note that will be preserved
 > by Prettier.

+### Links
+- **Accessibility:** Use descriptive anchor text; avoid "click here." Ensure the
+  link makes sense out of context, such as when being read by a screen reader.
+- **Use relative links in docs:** Use relative links in documentation (`/docs/`)
+  to ensure portability. Use paths relative to the current file's directory
+  (for example, `../tools/` from `docs/cli/`). Do not include the `/docs/`
+  section of a path, but do verify that the resulting relative link exists. This
+  does not apply to meta files such as README.MD and CONTRIBUTING.MD.
+- **When changing headings, check for deep links:** If a user is changing a
+  heading, check for deep links to that heading in other pages and update
+  accordingly.
+
 ### Structure
 - **BLUF:** Start with an introduction explaining what to expect.
 - **Experimental features:** If a feature is clearly noted as experimental,
@@ -157,7 +167,6 @@ documentation.
 - **Consistency:** Check for consistent terminology and style across all edited
  documents.

-
 ## Phase 4: Verification and finalization
 Perform a final quality check to ensure that all changes are correctly formatted
 and that all links are functional.
@@ -0,0 +1,69 @@
+---
+name: review-duplication
+description: Use this skill during code reviews to proactively investigate the codebase for duplicated functionality, reinvented wheels, or failure to reuse existing project best practices and shared utilities.
+---
+
+# Review Duplication
+
+## Overview
+
+This skill provides a structured workflow for investigating a codebase during a code review to identify duplicated logic, reinvented utilities, and missed opportunities to reuse established patterns. By executing this workflow, you ensure that new code integrates seamlessly with the existing project architecture.
+
+## Workflow: Investigating for Duplication
+
+When reviewing code, perform the following steps before finalizing your review:
+
+### 1. Extract Core Logic
+Analyze the new code to identify the core algorithms, utility functions, generic data structures, or UI components being introduced. Look beyond the specific business logic to see the underlying mechanics.
+
+### 2. Hypothesize Existing Locations & Trace Dependencies
+Think about where this type of code *would* live if it already existed in the project. Provide absolute paths from the repo root to disambiguate.
+- **Utilities:** `packages/core/src/utils/`, `packages/cli/src/utils/`
+- **UI Components:** `packages/cli/src/ui/components/`, `packages/cli/src/ui/`
+- **Services:** `packages/core/src/services/`, `packages/cli/src/services/`
+- **Configuration:** `packages/core/src/config/`, `packages/cli/src/config/`
+- **Core Logic:** Call out `packages/core/` if functionality does not appear React UI specific.
+
+**Trace Third-Party Dependencies:** If the PR introduces a new import for a utility library (e.g., `lodash.merge`, `date-fns`), trace how and where the project currently uses that library. There is likely an existing wrapper or shared utility.
+
+**Check Package Files:** Before flagging a custom implementation of a complex algorithm, check `package.json` to see if a standard library (like `lodash` or `uuid`) is already installed that provides this functionality.
+
+### 3. Investigate the Codebase (Sub-Agent Delegation)
+Delegate the heavy lifting of codebase investigation to specialized sub-agents. They are optimized to perform deep searches and semantic mapping without bloating your session history.
+
+To ensure a comprehensive review, you MUST formulate highly specific objectives for the sub-agents, providing them with the "scents" you discovered in Step 1.
+
+- **Codebase Investigator:** Use the `codebase_investigator` as your primary researcher. When delegating, formulate an objective that asks specific, investigative questions about the codebase, explicitly including these search vectors:
+  - **Structural Similarity:** Ask if existing code uses the same underlying APIs (e.g., "Does any existing code use `Intl.DateTimeFormat` or `setTimeout` for similar purposes?").
+  - **Naming Conventions:** Ask if there are existing symbols with similar naming patterns (e.g., "Are there existing symbols with naming patterns like `*Format*` or `*Debounce*`?").
+  - **Comments & Documentation:** Ask if keywords from the PR's comments or JSDoc exist in describing similar behavior elsewhere.
+  - **Architectural Fit:** Ask where this type of logic is currently centralized (e.g., "Where is centralized date formatting logic located?").
+  - **Refactoring Guidance:** Crucially, ask the sub-agent to explain *how* the new code could be refactored to use any existing logic it finds.
+- **Generalist Agent:** Use the `generalist` for detailed, turn-intensive comparisons. For example: "Review the implementation of `MyNewComponent` in the PR and compare it semantically against all components in `packages/ui/src`. Are there any existing components that could be extended or used instead?"
+- **Retain Fast Path for Simple Searches:** For extremely simple, unambiguous checks (e.g., "Does `package.json` include `lodash`?"), perform a direct search to save time. Default to delegation for any open-ended "investigations."
+
+### 4. Evaluate Best Practices
+Check if the new code aligns with the project's established conventions.
+- **Error Handling:** Does it use the project's standard error classes or logging mechanisms?
+- **State Management:** Does it bypass established stores or contexts?
+- **Styling:** Does it hardcode colors or spacing instead of using theme variables?
+If the PR introduces a new pattern, compare it against the documented standards and explicitly confirm if an existing project pattern should have been used instead.
+
+### 5. Formulate Constructive Feedback
+If you discover that the PR duplicates existing functionality or ignores a best practice:
+- Provide a clear review comment.
+- **Identify the Source:** Explicitly mention the absolute or project-relative file path and the specific symbol (function, component, class) that should be reused.
+- **Implementation Guidance:** Provide a brief code snippet or a clear explanation showing **how** to integrate the existing code to fulfill the task's requirements.
+- **Explain the Value:** Briefly explain why reusing the existing code is beneficial (e.g., maintainability, consistency, built-in edge case handling).
+
+Example comment:
+> "It looks like this PR introduces a new `formatDate` utility. We already have a robust, tested `formatDate` function in `src/utils/dateHelpers.ts`. 
+>
+> You can replace your implementation by importing it like this:
+> ```typescript
+> import { formatDate } from '../utils/dateHelpers';
+> 
+> // Then use it here:
+> const displayDate = formatDate(userDate, 'MMM Do, YYYY');
+> ```
+> Reusing this ensures that the date formatting remains consistent with the rest of the application and handles timezone conversions correctly."
@@ -22,6 +22,3 @@ Makefile eol=lf
 *.eot binary
 *.ttf binary
 *.otf binary
-
-# Configure GitHub to treat Markdoc files as Markdown
-*.mdoc linguist-language=Markdown
@@ -334,8 +334,20 @@ jobs:
        if: "${{ steps.check_evals.outputs.should_run == 'true' }}"
        env:
          GEMINI_API_KEY: '${{ secrets.GEMINI_API_KEY }}'
+          GEMINI_MODEL: 'gemini-3-pro-preview'
+          # Disable Vitest internal retries to avoid double-retrying;
+          # custom retry logic is handled in evals/test-helper.ts
+          VITEST_RETRY: 0
        run: 'npm run test:always_passing_evals'

+      - name: 'Upload Reliability Logs'
+        if: "always() && steps.check_evals.outputs.should_run == 'true'"
+        uses: 'actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02' # ratchet:actions/upload-artifact@v4
+        with:
+          name: 'eval-logs-${{ github.run_id }}-${{ github.run_attempt }}'
+          path: 'evals/logs/api-reliability.jsonl'
+          retention-days: 7
+
  e2e:
    name: 'E2E'
    if: |
@@ -61,6 +61,8 @@ jobs:
          GEMINI_MODEL: '${{ matrix.model }}'
          RUN_EVALS: "${{ github.event.inputs.run_all != 'false' }}"
          TEST_NAME_PATTERN: '${{ github.event.inputs.test_name_pattern }}'
+          # Disable Vitest internal retries to avoid double-retrying;
+          # custom retry logic is handled in evals/test-helper.ts
          VITEST_RETRY: 0
        run: |
          CMD="npm run test:all_evals"
@@ -30,7 +30,7 @@ Learn all about Gemini CLI in our [documentation](https://geminicli.com/docs/).
 ## 📦 Installation

 See
-[Gemini CLI installation, execution, and releases](https://geminicli.com/docs/get-started/installation/)
+[Gemini CLI installation, execution, and releases](./docs/get-started/installation.md)
 for recommended system specifications and a detailed installation guide.

 ### Quick Install
@@ -1,6 +1,6 @@
-# Latest stable release: v0.35.0
+# Latest stable release: v0.35.2

-Released: March 24, 2026
+Released: March 26, 2026

 For most users, our latest stable release is the recommended release. Install
 the latest stable version with:
@@ -29,6 +29,11 @@ npm install -g @google/gemini-cli

 ## What's Changed

+- fix(core): allow disabling environment variable redaction by @galz10 in
+  [#23927](https://github.com/google-gemini/gemini-cli/pull/23927)
+- fix(a2a-server): A2A server should execute ask policies in interactive mode by
+  @keith.schaab in
+  [#23831](https://github.com/google-gemini/gemini-cli/pull/23831)
 - feat(cli): customizable keyboard shortcuts by @scidomino in
  [#21945](https://github.com/google-gemini/gemini-cli/pull/21945)
 - feat(core): Thread `AgentLoopContext` through core. by @joshualitt in
@@ -380,4 +385,4 @@ npm install -g @google/gemini-cli
  [#23585](https://github.com/google-gemini/gemini-cli/pull/23585)

 **Full Changelog**:
-https://github.com/google-gemini/gemini-cli/compare/v0.34.0...v0.35.0
+https://github.com/google-gemini/gemini-cli/compare/v0.34.0...v0.35.2
@@ -1,6 +1,6 @@
-# Preview release: v0.36.0-preview.3
+# Preview release: v0.36.0-preview.5

-Released: March 25, 2026
+Released: March 27, 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).
@@ -31,6 +31,13 @@ npm install -g @google/gemini-cli@preview

 ## What's Changed

+- 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)
+- 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)
+- feat(core): support inline agentCardJson for remote agents by @adamfweidman in
+  [#23743](https://github.com/google-gemini/gemini-cli/pull/23743)
 - fix(patch): cherry-pick 055ff92 to release/v0.36.0-preview.0-pr-23672 to patch
  version v0.36.0-preview.0 and create version 0.36.0-preview.1 by
  @gemini-cli-robot in
@@ -379,4 +386,4 @@ npm install -g @google/gemini-cli@preview
  [#23666](https://github.com/google-gemini/gemini-cli/pull/23666)

 **Full Changelog**:
-https://github.com/google-gemini/gemini-cli/compare/v0.35.0-preview.5...v0.36.0-preview.3
+https://github.com/google-gemini/gemini-cli/compare/v0.35.0-preview.5...v0.36.0-preview.5
@@ -0,0 +1,126 @@
+# ACP Mode
+
+ACP (Agent Client Protocol) mode is a special operational mode of Gemini CLI
+designed for programmatic control, primarily for IDE and other developer tool
+integrations. It uses a JSON-RPC protocol over stdio to communicate between
+Gemini CLI agent and a client.
+
+To start Gemini CLI in ACP mode, use the `--acp` flag:
+
+```bash
+gemini --acp
+```
+
+## Agent Client Protocol (ACP)
+
+ACP is an open protocol that standardizes how AI coding agents communicate with
+code editors and IDEs. It addresses the challenge of fragmented distribution,
+where agents traditionally needed custom integrations for each client. With ACP,
+developers can implement their agent once, and it becomes compatible with any
+ACP-compliant editor.
+
+For a comprehensive introduction to ACP, including its architecture and
+benefits, refer to the official
+[ACP Introduction](https://agentclientprotocol.com/get-started/introduction)
+documentation.
+
+### Existing integrations using ACP
+
+The ACP Agent Registry simplifies the distribution and management of
+ACP-compatible agents across various IDEs. Gemini CLI is an ACP-compatible agent
+and can be found in this registry.
+
+For more general information about the registry, and how to use it with specific
+IDEs like JetBrains and Zed, refer to the
+[IDE Integration](../ide-integration/index.md) documentation.
+
+You can also find more information on the official
+[ACP Agent Registry](https://agentclientprotocol.com/get-started/registry) page.
+
+## Architecture and protocol basics
+
+ACP mode establishes a client-server relationship between your tool (the client)
+and Gemini CLI (the server).
+
+- **Communication:** The entire communication happens over standard input/output
+  (stdio) using the JSON-RPC 2.0 protocol.
+- **Client's role:** The client is responsible for sending requests (e.g.,
+  prompts) and handling responses and notifications from Gemini CLI.
+- **Gemini CLI's role:** In ACP mode, Gemini CLI listens for incoming JSON-RPC
+  requests, processes them, and sends back responses.
+
+The core of the ACP implementation can be found in
+`packages/cli/src/acp/acpClient.ts`.
+
+### Extending with MCP
+
+ACP can be used with the Model Context Protocol (MCP). This lets an ACP client
+(like an IDE) expose its own functionality as "tools" that the Gemini model can
+use.
+
+1.  The client implements an **MCP server** that advertises its tools.
+2.  During the ACP `initialize` handshake, the client provides the connection
+    details for its MCP server.
+3.  Gemini CLI connects to the MCP server, discovers the available tools, and
+    makes them available to the AI model.
+4.  When the model decides to use one of these tools, Gemini CLI sends a tool
+    call request to the MCP server.
+
+This mechanism lets for a powerful, two-way integration where the agent can
+leverage the IDE's capabilities to perform tasks. The MCP client logic is in
+`packages/core/src/tools/mcp-client.ts`.
+
+## Capabilities and supported methods
+
+The ACP protocol exposes a number of methods for ACP clients (e.g. IDEs) to
+control Gemini CLI.
+
+### Core methods
+
+- `initialize`: Establishes the initial connection and lets the client to
+  register its MCP server.
+- `authenticate`: Authenticates the user.
+- `newSession`: Starts a new chat session.
+- `loadSession`: Loads a previous session.
+- `prompt`: Sends a prompt to the agent.
+- `cancel`: Cancels an ongoing prompt.
+
+### Session control
+
+- `setSessionMode`: Allows changing the approval level for tool calls (e.g., to
+  `auto-approve`).
+- `unstable_setSessionModel`: Changes the model for the current session.
+
+### File system proxy
+
+ACP includes a proxied file system service. This means that when the agent needs
+to read or write files, it does so through the ACP client. This is a security
+feature that ensures the agent only has access to the files that the client (and
+by extension, the user) has explicitly allowed.
+
+## Debugging and telemetry
+
+You can get insights into the ACP communication and the agent's behavior through
+debugging logs and telemetry.
+
+### Debugging logs
+
+To enable general debugging logs, start Gemini CLI with the `--debug` flag:
+
+```bash
+gemini --acp --debug
+```
+
+### Telemetry
+
+For more detailed telemetry, you can use the following environment variables to
+capture telemetry data to a file:
+
+- `GEMINI_TELEMETRY_ENABLED=true`
+- `GEMINI_TELEMETRY_TARGET=local`
+- `GEMINI_TELEMETRY_OUTFILE=/path/to/your/log.json`
+
+This will write a JSON log file containing detailed information about all the
+events happening within the agent, including ACP requests and responses. The
+integration test `integration-tests/acp-telemetry.test.ts` provides a working
+example of how to set this up.
@@ -15,14 +15,14 @@ CLI works in the background.

 ## Requirements

-Currently, system notifications are only supported on macOS.
-
 ### Terminal support

 The CLI uses the OSC 9 terminal escape sequence to trigger system notifications.
-This is supported by several modern terminal emulators. If your terminal does
-not support OSC 9 notifications, Gemini CLI falls back to a system alert sound
-to get your attention.
+This is supported by several modern terminal emulators including iTerm2,
+WezTerm, Ghostty, and Kitty. If your terminal does not support OSC 9
+notifications, Gemini CLI falls back to a terminal bell (BEL) to get your
+attention. Most terminals respond to BEL with a taskbar flash or system alert
+sound.

 ## Enable notifications

@@ -39,7 +39,9 @@ To start Plan Mode while using Gemini CLI:
  the rotation when Gemini CLI is actively processing or showing confirmation
  dialogs.

- **Command:** Type `/plan` in the input box.
+- **Command:** Type `/plan [goal]` in the input box. The `[goal]` is optional;
+  for example, `/plan implement authentication` will switch to Plan Mode and
+  immediately submit the prompt to the model.

 - **Natural Language:** Ask Gemini CLI to "start a plan for...". Gemini CLI
  calls the
@@ -29,7 +29,7 @@ they appear in the UI.
 | Vim Mode                | `general.vimMode`                  | Enable Vim keybindings                                                                                                                                                                                                                                        | `false`     |
 | Default Approval Mode   | `general.defaultApprovalMode`      | The default approval mode for tool execution. 'default' prompts for approval, 'auto_edit' auto-approves edit tools, and 'plan' is read-only mode. YOLO mode (auto-approve all actions) can only be enabled via command line (--yolo or --approval-mode=yolo). | `"default"` |
 | Enable Auto Update      | `general.enableAutoUpdate`         | Enable automatic updates.                                                                                                                                                                                                                                     | `true`      |
-| Enable Notifications    | `general.enableNotifications`      | Enable run-event notifications for action-required prompts and session completion. Currently macOS only.                                                                                                                                                      | `false`     |
+| Enable Notifications    | `general.enableNotifications`      | Enable run-event notifications for action-required prompts and session completion.                                                                                                                                                                            | `false`     |
 | Plan Directory          | `general.plan.directory`           | The directory where planning artifacts are stored. If not specified, defaults to the system temporary directory. A custom directory requires a policy to allow write access in Plan Mode.                                                                     | `undefined` |
 | Plan Model Routing      | `general.plan.modelRouting`        | Automatically switch between Pro and Flash models based on Plan Mode status. Uses Pro for the planning phase and Flash for the implementation phase.                                                                                                          | `true`      |
 | Retry Fetch Errors      | `general.retryFetchErrors`         | Retry on "exception TypeError: fetch failed sending request" errors.                                                                                                                                                                                          | `true`      |
@@ -155,17 +155,21 @@ they appear in the UI.

 ### Experimental

-| UI Label                   | Setting                                  | Description                                                                                                                                               | Default |
-| -------------------------- | ---------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------- | ------- |
-| Enable Tool Output Masking | `experimental.toolOutputMasking.enabled` | Enables tool output masking to save tokens.                                                                                                               | `true`  |
-| Enable Git Worktrees       | `experimental.worktrees`                 | Enable automated Git worktree management for parallel work.                                                                                               | `false` |
-| Use OSC 52 Paste           | `experimental.useOSC52Paste`             | Use OSC 52 for pasting. This may be more robust than the default system when using remote terminal sessions (if your terminal is configured to allow it). | `false` |
-| Use OSC 52 Copy            | `experimental.useOSC52Copy`              | Use OSC 52 for copying. This may be more robust than the default system when using remote terminal sessions (if your terminal is configured to allow it). | `false` |
-| Plan                       | `experimental.plan`                      | Enable Plan Mode.                                                                                                                                         | `true`  |
-| Model Steering             | `experimental.modelSteering`             | Enable model steering (user hints) to guide the model during tool execution.                                                                              | `false` |
-| Direct Web Fetch           | `experimental.directWebFetch`            | Enable web fetch behavior that bypasses LLM summarization.                                                                                                | `false` |
-| Memory Manager Agent       | `experimental.memoryManager`             | Replace the built-in save_memory tool with a memory manager subagent that supports adding, removing, de-duplicating, and organizing memories.             | `false` |
-| Topic & Update Narration   | `experimental.topicUpdateNarration`      | Enable the experimental Topic & Update communication model for reduced chattiness and structured progress reporting.                                      | `false` |
+| UI Label                           | Setting                                        | Description                                                                                                                                               | Default |
+| ---------------------------------- | ---------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------- | ------- |
+| Enable Tool Output Masking         | `experimental.toolOutputMasking.enabled`       | Enables tool output masking to save tokens.                                                                                                               | `true`  |
+| Enable Git Worktrees               | `experimental.worktrees`                       | Enable automated Git worktree management for parallel work.                                                                                               | `false` |
+| Use OSC 52 Paste                   | `experimental.useOSC52Paste`                   | Use OSC 52 for pasting. This may be more robust than the default system when using remote terminal sessions (if your terminal is configured to allow it). | `false` |
+| Use OSC 52 Copy                    | `experimental.useOSC52Copy`                    | Use OSC 52 for copying. This may be more robust than the default system when using remote terminal sessions (if your terminal is configured to allow it). | `false` |
+| Plan                               | `experimental.plan`                            | Enable Plan Mode.                                                                                                                                         | `true`  |
+| Model Steering                     | `experimental.modelSteering`                   | Enable model steering (user hints) to guide the model during tool execution.                                                                              | `false` |
+| Direct Web Fetch                   | `experimental.directWebFetch`                  | Enable web fetch behavior that bypasses LLM summarization.                                                                                                | `false` |
+| Memory Manager Agent               | `experimental.memoryManager`                   | Replace the built-in save_memory tool with a memory manager subagent that supports adding, removing, de-duplicating, and organizing memories.             | `false` |
+| Agent History Truncation           | `experimental.agentHistoryTruncation`          | Enable truncation window logic for the Agent History Provider.                                                                                            | `false` |
+| Agent History Truncation Threshold | `experimental.agentHistoryTruncationThreshold` | The maximum number of messages before history is truncated.                                                                                               | `30`    |
+| Agent History Retained Messages    | `experimental.agentHistoryRetainedMessages`    | The number of recent messages to retain after truncation.                                                                                                 | `15`    |
+| Agent History Summarization        | `experimental.agentHistorySummarization`       | Enable summarization of truncated content via a small model for the Agent History Provider.                                                               | `false` |
+| Topic & Update Narration           | `experimental.topicUpdateNarration`            | Enable the experimental Topic & Update communication model for reduced chattiness and structured progress reporting.                                      | `false` |

 ### Skills

@@ -1,141 +0,0 @@
-# Gemini CLI examples
-
-Gemini CLI helps you automate common engineering tasks by combining AI reasoning
-with local system tools. This document provides examples of how to use the CLI
-for file management, code analysis, and data transformation.
-
-<!-- prettier-ignore -->
-> [!NOTE]
-> These examples demonstrate potential capabilities. Your actual
-> results can vary based on the model used and your project environment.
-
-## Rename your photographs based on content
-
-You can use Gemini CLI to automate file management tasks that require visual
-analysis. In this example, Gemini CLI renames images based on their actual
-subject matter.
-
-Scenario: You have a folder containing the following files:
-
-```bash
-photos/photo1.png
-photos/photo2.png
-photos/photo3.png
-```
-
-Give Gemini the following prompt:
-
-```cli
-Rename the photos in my "photos" directory based on their contents.
-```
-
-Result: Gemini asks for permission to rename your files.
-
-Select **Allow once** and your files are renamed:
-
-```bash
-photos/yellow_flowers.png
-photos/antique_dresser.png
-photos/green_android_robot.png
-```
-
-## Explain a repository by reading its code
-
-Gemini CLI is effective for rapid codebase exploration. The following example
-shows how to ask Gemini CLI to fetch, analyze, and summarize a remote project.
-
-Scenario: You want to understand how a popular open-source utility works by
-inspecting its code, not just its README.
-
-Give Gemini CLI the following prompt:
-
-```cli
-Clone the 'chalk' repository from https://github.com/chalk/chalk, read its key source files, and explain how it works.
-```
-
-Result: Gemini performs a sequence of actions to answer your request.
-
-1.  First, it asks for permission to run `git clone` to download the repository.
-2.  Next, it finds the important source files and asks for permission to read
-    them.
-3.  Finally, after analyzing the code, it provides a summary.
-
-Gemini CLI returns an explanation based on the actual source code:
-
-```markdown
-The `chalk` library is a popular npm package for styling terminal output with
-colors. After analyzing the source code, here's how it works:
-
- **Core functionality:** The main file sets up a chainable API. Each color or
-  modifier (like `bold` or `italic`) is a getter that appends the corresponding
-  ANSI escape code to an internal stack.
-
- **Chainable API:** When you call `chalk.red.bold('Hello')`, you are accessing
-  getters. The `red` getter adds the red color code, and the `bold` getter adds
-  the bold code.
-
- **Output generation:** When the chain is treated as a string (e.g., in
-  `console.log`), a final `toString()` method is called. This method joins all
-  the stored ANSI codes, wraps them around the input string ('Hello'), and adds
-  a reset code at the end. This produces the final, styled string that the
-  terminal can render.
-```
-
-## Combine two spreadsheets into one spreadsheet
-
-Gemini CLI can process and transform data across multiple files. Use this
-capability to merge reports or reformat data sets without manual copying.
-
-Scenario: You have two .csv files: `Revenue - 2023.csv` and
-`Revenue - 2024.csv`. Each file contains monthly revenue figures.
-
-Give Gemini CLI the following prompt:
-
-```cli
-Combine the two .csv files into a single .csv file, with each year a different column.
-```
-
-Result: Gemini CLI reads each file and then asks for permission to write a new
-file. Provide your permission and Gemini CLI provides the combined data:
-
-```csv
-Month,2023,2024
-January,0,1000
-February,0,1200
-March,0,2400
-April,900,500
-May,1000,800
-June,1000,900
-July,1200,1000
-August,1800,400
-September,2000,2000
-October,2400,3400
-November,3400,1800
-December,2100,9000
-```
-
-## Run unit tests
-
-Gemini CLI can generate boilerplate code and tests based on your existing
-implementation. This example demonstrates how to request code coverage for a
-JavaScript component.
-
-Scenario: You've written a simple login page. You wish to write unit tests to
-ensure that your login page has code coverage.
-
-Give Gemini CLI the following prompt:
-
-```cli
-Write unit tests for Login.js.
-```
-
-Result: Gemini CLI asks for permission to write a new file and creates a test
-for your login page.
-
-## Next steps
-
- Follow the [File management](../cli/tutorials/file-management.md) guide to
-  start working with your codebase.
- Follow the [Quickstart](./index.md) to start your first session.
- See the [Cheatsheet](../cli/cli-reference.md) for a quick reference of
-  available commands.
@@ -1,6 +1,6 @@
 # Gemini 3 Pro and Gemini 3 Flash on Gemini CLI

-Gemini 3 Pro and Gemini 3 Flash are available on Gemini CLI for all users!
+Learn about how you can use Gemini 3 Pro and Gemini 3 Flash on Gemini CLI.

 <!-- prettier-ignore -->
 > [!NOTE]
@@ -24,7 +24,7 @@ Once Gemini CLI is installed, run Gemini CLI from your command line:
 gemini
 ```

-For more installation options, see [Gemini CLI Installation](./installation/).
+For more installation options, see [Gemini CLI Installation](./installation.md).

 ## Authenticate

@@ -62,7 +62,133 @@ Once installed and authenticated, you can start using Gemini CLI by issuing
 commands and prompts in your terminal. Ask it to generate code, explain files,
 and more.

-To explore the power of Gemini CLI, see [Gemini CLI examples](./examples.md).
+<!-- prettier-ignore -->
+> [!NOTE]
+> These examples demonstrate potential capabilities. Your actual
+> results can vary based on the model used and your project environment.
+
+### Rename your photographs based on content
+
+You can use Gemini CLI to automate file management tasks that require visual
+analysis. In this example, Gemini CLI renames images based on their actual
+subject matter.
+
+Scenario: You have a folder containing the following files:
+
+```bash
+photos/photo1.png
+photos/photo2.png
+photos/photo3.png
+```
+
+Give Gemini the following prompt:
+
+```cli
+Rename the photos in my "photos" directory based on their contents.
+```
+
+Result: Gemini asks for permission to rename your files.
+
+Select **Allow once** and your files are renamed:
+
+```bash
+photos/yellow_flowers.png
+photos/antique_dresser.png
+photos/green_android_robot.png
+```
+
+### Explain a repository by reading its code
+
+Gemini CLI is effective for rapid codebase exploration. The following example
+shows how to ask Gemini CLI to fetch, analyze, and summarize a remote project.
+
+Scenario: You want to understand how a popular open-source utility works by
+inspecting its code, not just its README.
+
+Give Gemini CLI the following prompt:
+
+```cli
+Clone the 'chalk' repository from https://github.com/chalk/chalk, read its key source files, and explain how it works.
+```
+
+Result: Gemini performs a sequence of actions to answer your request.
+
+1.  First, it asks for permission to run `git clone` to download the repository.
+2.  Next, it finds the important source files and asks for permission to read
+    them.
+3.  Finally, after analyzing the code, it provides a summary.
+
+Gemini CLI returns an explanation based on the actual source code:
+
+```markdown
+The `chalk` library is a popular npm package for styling terminal output with
+colors. After analyzing the source code, here's how it works:
+
+- **Core functionality:** The main file sets up a chainable API. Each color or
+  modifier (like `bold` or `italic`) is a getter that appends the corresponding
+  ANSI escape code to an internal stack.
+
+- **Chainable API:** When you call `chalk.red.bold('Hello')`, you are accessing
+  getters. The `red` getter adds the red color code, and the `bold` getter adds
+  the bold code.
+
+- **Output generation:** When the chain is treated as a string (e.g., in
+  `console.log`), a final `toString()` method is called. This method joins all
+  the stored ANSI codes, wraps them around the input string ('Hello'), and adds
+  a reset code at the end. This produces the final, styled string that the
+  terminal can render.
+```
+
+### Combine two spreadsheets into one spreadsheet
+
+Gemini CLI can process and transform data across multiple files. Use this
+capability to merge reports or reformat data sets without manual copying.
+
+Scenario: You have two .csv files: `Revenue - 2023.csv` and
+`Revenue - 2024.csv`. Each file contains monthly revenue figures.
+
+Give Gemini CLI the following prompt:
+
+```cli
+Combine the two .csv files into a single .csv file, with each year a different column.
+```
+
+Result: Gemini CLI reads each file and then asks for permission to write a new
+file. Provide your permission and Gemini CLI provides the combined data:
+
+```csv
+Month,2023,2024
+January,0,1000
+February,0,1200
+March,0,2400
+April,900,500
+May,1000,800
+June,1000,900
+July,1200,1000
+August,1800,400
+September,2000,2000
+October,2400,3400
+November,3400,1800
+December,2100,9000
+```
+
+### Run unit tests
+
+Gemini CLI can generate boilerplate code and tests based on your existing
+implementation. This example demonstrates how to request code coverage for a
+JavaScript component.
+
+Scenario: You've written a simple login page. You wish to write unit tests to
+ensure that your login page has code coverage.
+
+Give Gemini CLI the following prompt:
+
+```cli
+Write unit tests for Login.js.
+```
+
+Result: Gemini CLI asks for permission to write a new file and creates a test
+for your login page.

 ## Check usage and quota

@@ -23,33 +23,34 @@ installation methods, and release types.
 We recommend most users install Gemini CLI using one of the following
 installation methods:

-{% tabs %}
-{% tabitem label="npm" %}
-Install globally with npm:
+- npm
+- Homebrew
+- MacPorts
+- Anaconda
+
+Note that Gemini CLI comes pre-installed on
+[**Cloud Shell**](https://docs.cloud.google.com/shell/docs) and
+[**Cloud Workstations**](https://cloud.google.com/workstations).
+
+### Install globally with npm

 ```bash
 npm install -g @google/gemini-cli
 ```
-{% /tabitem %}

-{% tabitem label="Homebrew" %}
-Install globally with Homebrew (macOS/Linux):
+### Install globally with Homebrew (macOS/Linux)

 ```bash
 brew install gemini-cli
 ```
-{% /tabitem %}

-{% tabitem label="MacPorts" %}
-Install globally with MacPorts (macOS):
+### Install globally with MacPorts (macOS)

 ```bash
 sudo port install gemini-cli
 ```
-{% /tabitem %}

-{% tabitem label="Anaconda" %}
-Install with Anaconda (for restricted environments):
+### Install with Anaconda (for restricted environments)

 ```bash
 # Create and activate a new environment
@@ -59,12 +60,6 @@ conda activate gemini_env
 # Install Gemini CLI globally via npm (inside the environment)
 npm install -g @google/gemini-cli
 ```
-{% /tabitem %}
-{% /tabs %}
-
-Note that Gemini CLI comes pre-installed on
-[**Cloud Shell**](https://docs.cloud.google.com/shell/docs) and
-[**Cloud Workstations**](https://cloud.google.com/workstations).

 ## Run Gemini CLI

@@ -107,7 +102,7 @@ the default way that the CLI executes tools that might have side effects.
  to run the CLI.
  ```bash
  # Run the published sandbox image
-  docker run --rm -it us-docker.pkg.dev/gemini-code-dev/gemini-cli/sandbox:latest
+  docker run --rm -it us-docker.pkg.dev/gemini-code-dev/gemini-cli/sandbox:0.1.1
  ```
 - **Using the `--sandbox` flag:** If you have Gemini CLI installed locally
  (using the standard installation described above), you can instruct it to run
@@ -1,157 +0,0 @@
-# Gemini CLI installation, execution, and releases
-
-This document provides an overview of Gemini CLI's system requirements,
-installation methods, and release types.
-
-## Recommended system specifications
-
- **Operating System:**
-  - macOS 15+
-  - Windows 11 24H2+
-  - Ubuntu 20.04+
- **Hardware:**
-  - "Casual" usage: 4GB+ RAM (short sessions, common tasks and edits)
-  - "Power" usage: 16GB+ RAM (long sessions, large codebases, deep context)
- **Runtime:** Node.js 20.0.0+
- **Shell:** Bash, Zsh, or PowerShell
- **Location:**
-  [Gemini Code Assist supported locations](https://developers.google.com/gemini-code-assist/resources/available-locations#americas)
- **Internet connection required**
-
-## Install Gemini CLI
-
-We recommend most users install Gemini CLI using one of the following
-installation methods:
-
-<Tabs>
-  <TabItem label="npm">
-  Install globally with npm:
-
-  ```bash
-  npm install -g @google/gemini-cli
-  ```
-  </TabItem>
-  <TabItem label="Homebrew">
-  Install globally with Homebrew (macOS/Linux):
-
-  ```bash
-  brew install gemini-cli
-  ```
-  </TabItem>
-</Tabs>
-
-Note that Gemini CLI comes pre-installed on
-[**Cloud Shell**](https://docs.cloud.google.com/shell/docs) and
-[**Cloud Workstations**](https://cloud.google.com/workstations).
-
-## Run Gemini CLI
-
-For most users, we recommend running Gemini CLI with the `gemini` command:
-
-```bash
-gemini
-```
-
-For a list of options and additional commands, see the
-[CLI cheatsheet](../cli/cli-reference.md).
-
-You can also run Gemini CLI using one of the following advanced methods:
-
- Run instantly with npx. You can run Gemini CLI without permanent installation.
- In a sandbox. This method offers increased security and isolation.
- From the source. This is recommended for contributors to the project.
-
-### Run instantly with npx
-
-```bash
-# Using npx (no installation required)
-npx @google/gemini-cli
-```
-
-You can also execute the CLI directly from the main branch on GitHub, which is
-helpful for testing features still in development:
-
-```bash
-npx https://github.com/google-gemini/gemini-cli
-```
-
-### Run in a sandbox (Docker/Podman)
-
-For security and isolation, Gemini CLI can be run inside a container. This is
-the default way that the CLI executes tools that might have side effects.
-
- **Directly from the registry:** You can run the published sandbox image
-  directly. This is useful for environments where you only have Docker and want
-  to run the CLI.
-  ```bash
-  # Run the published sandbox image
-  docker run --rm -it us-docker.pkg.dev/gemini-code-dev/gemini-cli/sandbox:latest
-  ```
- **Using the `--sandbox` flag:** If you have Gemini CLI installed locally
-  (using the standard installation described above), you can instruct it to run
-  inside the sandbox container.
-  ```bash
-  gemini --sandbox -y -p "your prompt here"
-  ```
-
-### Run from source (recommended for Gemini CLI contributors)
-
-Contributors to the project will want to run the CLI directly from the source
-code.
-
- **Development mode:** This method provides hot-reloading and is useful for
-  active development.
-  ```bash
-  # From the root of the repository
-  npm run start
-  ```
- **Production-like mode (linked package):** This method simulates a global
-  installation by linking your local package. It's useful for testing a local
-  build in a production workflow.
-
-  ```bash
-  # Link the local cli package to your global node_modules
-  npm link packages/cli
-
-  # Now you can run your local version using the `gemini` command
-  gemini
-  ```
-
-## Releases
-
-Gemini CLI has three release channels: nightly, preview, and stable. For most
-users, we recommend the stable release, which is the default installation.
-
-### Stable
-
-New stable releases are published each week. The stable release is the promotion
-of last week's `preview` release along with any bug fixes. The stable release
-uses `latest` tag, but omitting the tag also installs the latest stable release
-by default:
-
-```bash
-# Both commands install the latest stable release.
-npm install -g @google/gemini-cli
-npm install -g @google/gemini-cli@latest
-```
-
-### Preview
-
-New preview releases will be published each week. These releases are not fully
-vetted and may contain regressions or other outstanding issues. Try out the
-preview release by using the `preview` tag:
-
-```bash
-npm install -g @google/gemini-cli@preview
-```
-
-### Nightly
-
-Nightly releases are published every day. The nightly release includes all
-changes from the main branch at time of release. It should be assumed there are
-pending validations and issues. You can help test the latest changes by
-installing with the `nightly` tag:
-
-```bash
-npm install -g @google/gemini-cli@nightly
-```
@@ -1,15 +1,29 @@
-# IDE integration
+# IDE Integration

 Gemini CLI can integrate with your IDE to provide a more seamless and
 context-aware experience. This integration allows the CLI to understand your
 workspace better and enables powerful features like native in-editor diffing.

-Currently, the supported IDEs are [Antigravity](https://antigravity.google),
-[Visual Studio Code](https://code.visualstudio.com/), and other editors that
-support VS Code extensions. To build support for other editors, see the
-[IDE Companion Extension Spec](./ide-companion-spec.md).
+There are two primary ways to integrate Gemini CLI with an IDE:

-## Features
+1.  **VS Code companion extension**: Install the "Gemini CLI Companion"
+    extension on [Antigravity](https://antigravity.google),
+    [Visual Studio Code](https://code.visualstudio.com/), or other VS Code
+    compatible editors.
+2.  **Agent Client Protocol (ACP)**: An open protocol for interoperability
+    between AI coding agents and IDEs. This method is used for integrations with
+    tools like JetBrains and Zed, which leverage the ACP Agent Registry for easy
+    discovery and installation of compatible agents like Gemini CLI.
+
+## VS Code companion extension
+
+The **Gemini CLI Companion extension** grants Gemini CLI direct access to your
+VS Code compatible IDEs and improves your experience by providing real-time
+context such as open files, cursor positions, and text selection. The extension
+also enables a native diffing interface so you can seamlessly review and apply
+AI-generated code changes directly within your editor.
+
+### Features

 - **Workspace context:** The CLI automatically gains awareness of your workspace
  to provide more relevant and accurate responses. This context includes:
@@ -19,8 +33,8 @@ support VS Code extensions. To build support for other editors, see the
    truncated).

 - **Native diffing:** When Gemini suggests code modifications, you can view the
-  changes directly within your IDE's native diff viewer. This allows you to
-  review, edit, and accept or reject the suggested changes seamlessly.
+  changes directly within your IDE's native diff viewer. This lets you review,
+  edit, and accept or reject the suggested changes seamlessly.

 - **VS Code commands:** You can access Gemini CLI features directly from the VS
  Code Command Palette (`Cmd+Shift+P` or `Ctrl+Shift+P`):
@@ -32,18 +46,18 @@ support VS Code extensions. To build support for other editors, see the
  - `Gemini CLI: View Third-Party Notices`: Displays the third-party notices for
    the extension.

-## Installation and setup
+### Installation and setup

 There are three ways to set up the IDE integration:

-### 1. Automatic nudge (recommended)
+#### 1. Automatic nudge (recommended)

 When you run Gemini CLI inside a supported editor, it will automatically detect
 your environment and prompt you to connect. Answering "Yes" will automatically
 run the necessary setup, which includes installing the companion extension and
 enabling the connection.

-### 2. Manual installation from CLI
+#### 2. Manual installation from CLI

 If you previously dismissed the prompt or want to install the extension
 manually, you can run the following command inside Gemini CLI:
@@ -54,7 +68,7 @@ manually, you can run the following command inside Gemini CLI:

 This will find the correct extension for your IDE and install it.

-### 3. Manual installation from a marketplace
+#### 3. Manual installation from a marketplace

 You can also install the extension directly from a marketplace.

@@ -75,9 +89,9 @@ You can also install the extension directly from a marketplace.
 > After manually installing the extension, you must run `/ide enable` in the CLI
 > to activate the integration.

-## Usage
+### Usage

-### Enabling and disabling
+#### Enabling and disabling

 You can control the IDE integration from within the CLI:

@@ -93,7 +107,7 @@ You can control the IDE integration from within the CLI:
 When enabled, Gemini CLI will automatically attempt to connect to the IDE
 companion extension.

-### Checking the status
+#### Checking the status

 To check the connection status and see the context the CLI has received from the
 IDE, run:
@@ -108,9 +122,9 @@ recently opened files it is aware of.
 <!-- prettier-ignore -->
 > [!NOTE]
 > The file list is limited to 10 recently accessed files within your
-> workspace and only includes local files on disk.)
+> workspace and only includes local files on disk.

-### Working with diffs
+#### Working with diffs

 When you ask Gemini to modify a file, it can open a diff view directly in your
 editor.
@@ -135,6 +149,63 @@ accepting them.
 If you select ‘Allow for this session’ in the CLI, changes will no longer show
 up in the IDE as they will be auto-accepted.

+## Agent Client Protocol (ACP)
+
+ACP is an open protocol that standardizes how AI coding agents communicate with
+code editors and IDEs. It addresses the challenge of fragmented distribution,
+where agents traditionally needed custom integrations for each client. With ACP,
+developers can implement their agent once, and it becomes compatible with any
+ACP-compliant editor.
+
+For a comprehensive introduction to ACP, including its architecture and
+benefits, refer to the official
+[ACP Introduction](https://agentclientprotocol.com/get-started/introduction)
+documentation.
+
+### The ACP Agent Registry
+
+Gemini CLI is officially available in the **ACP Agent Registry**. This allows
+you to install and update Gemini CLI directly within supporting IDEs and
+eliminates the need for manual downloads or IDE-specific extensions.
+
+Using the registry ensures:
+
+- **Ease of use**: Discover and install agents directly within your IDE
+  settings.
+- **Latest versions**: Ensures users always have access to the most up-to-date
+  agent implementations.
+
+For more details on how the registry works, visit the official
+[ACP Agent Registry](https://agentclientprotocol.com/get-started/registry) page.
+You can learn about how specific IDEs leverage this integration in the following
+section.
+
+### IDE-specific integration
+
+Gemini CLI is an ACP-compatible agent available in the ACP Agent Registry.
+Here’s how different IDEs leverage the ACP and the registry:
+
+#### JetBrains IDEs
+
+JetBrains IDEs (like IntelliJ IDEA, PyCharm, or GoLand) offer built-in registry
+support, allowing users to find and install ACP-compatible agents directly.
+
+For more details, refer to the official
+[JetBrains AI Blog announcement](https://blog.jetbrains.com/ai/2026/01/acp-agent-registry/).
+
+#### Zed
+
+Zed, a modern code editor, also integrates with the ACP Agent Registry. This
+allows Zed users to easily browse, install, and manage ACP agents.
+
+Learn more about Zed's integration with the ACP Registry in their
+[blog post](https://zed.dev/blog/acp-registry).
+
+#### Other ACP-compatible IDEs
+
+Any other IDE that supports the ACP Agent Registry can install Gemini CLI
+directly through their in-built registry features.
+
 ## Using with sandboxing

 If you are using Gemini CLI within a sandbox, please be aware of the following:
@@ -151,10 +222,9 @@ If you are using Gemini CLI within a sandbox, please be aware of the following:

 ## Troubleshooting

-If you encounter issues with IDE integration, here are some common error
-messages and how to resolve them.
+### VS Code companion extension errors

-### Connection errors
+#### Connection errors

 - **Message:**
  `🔴 Disconnected: Failed to connect to IDE companion extension in [IDE Name]. Please ensure the extension is running. To install the extension, run /ide install.`
@@ -174,7 +244,7 @@ messages and how to resolve them.
  - **Solution:** Run `/ide enable` to try and reconnect. If the issue
    continues, open a new terminal window or restart your IDE.

-### Manual PID override
+#### Manual PID override

 If automatic IDE detection fails, or if you are running Gemini CLI in a
 standalone terminal and want to manually associate it with a specific IDE
@@ -196,7 +266,7 @@ $env:GEMINI_CLI_IDE_PID=12345
 When this variable is set, Gemini CLI will skip automatic detection and attempt
 to connect using the provided PID.

-### Configuration errors
+#### Configuration errors

 - **Message:**
  `🔴 Disconnected: Directory mismatch. Gemini CLI is running in a different location than the open workspace in [IDE Name]. Please run the CLI from one of the following directories: [List of directories]`
@@ -210,7 +280,7 @@ to connect using the provided PID.
  - **Cause:** You have no workspace open in your IDE.
  - **Solution:** Open a workspace in your IDE and restart the CLI.

-### General errors
+#### General errors

 - **Message:**
  `IDE integration is not supported in your current environment. To use this feature, run Gemini CLI in one of these supported IDEs: [List of IDEs]`
@@ -220,9 +290,14 @@ to connect using the provided PID.
    IDE, like Antigravity or VS Code.

 - **Message:**
-  `No installer is available for IDE. Please install the Gemini CLI Companion extension manually from the marketplace.`
+  `No installer is available for IDE. Please install Gemini CLI Companion extension manually from the marketplace.`
  - **Cause:** You ran `/ide install`, but the CLI does not have an automated
    installer for your specific IDE.
  - **Solution:** Open your IDE's extension marketplace, search for "Gemini CLI
    Companion", and
    [install it manually](#3-manual-installation-from-a-marketplace).
+
+### ACP integration errors
+
+For issues related to ACP integration, please refer to the debugging and
+telemetry section in the [ACP Mode](../cli/acp-mode.md) documentation.
@@ -15,12 +15,10 @@ npm install -g @google/gemini-cli
 Jump in to Gemini CLI.

 - **[Quickstart](./get-started/index.md):** Your first session with Gemini CLI.
- **[Installation](./get-started/installation/):** How to install Gemini CLI on
-  your system.
+- **[Installation](./get-started/installation.md):** How to install Gemini CLI
+  on your system.
 - **[Authentication](./get-started/authentication.md):** Setup instructions for
  personal and enterprise accounts.
- **[Examples](./get-started/examples.md):** Practical examples of Gemini CLI in
-  action.
 - **[CLI cheatsheet](./cli/cli-reference.md):** A quick reference for common
  commands and options.
 - **[Gemini 3 on Gemini CLI](./get-started/gemini-3.md):** Learn about Gemini 3
@@ -13,6 +13,7 @@
  "/docs/faq": "/docs/resources/faq",
  "/docs/get-started/configuration": "/docs/reference/configuration",
  "/docs/get-started/configuration-v1": "/docs/reference/configuration",
+  "/docs/get-started/examples": "/docs/get-started/index",
  "/docs/index": "/docs",
  "/docs/quota-and-pricing": "/docs/resources/quota-and-pricing",
  "/docs/tos-privacy": "/docs/resources/tos-privacy",
@@ -133,7 +133,7 @@ their corresponding top-level category object in your `settings.json` file.

 - **`general.enableNotifications`** (boolean):
  - **Description:** Enable run-event notifications for action-required prompts
-    and session completion. Currently macOS only.
+    and session completion.
  - **Default:** `false`

 - **`general.checkpointing.enabled`** (boolean):
@@ -670,6 +670,11 @@ their corresponding top-level category object in your `settings.json` file.
        "modelConfig": {
          "model": "gemini-3-pro-preview"
        }
+      },
+      "agent-history-provider-summarizer": {
+        "modelConfig": {
+          "model": "gemini-3-flash-preview"
+        }
      }
    }
    ```
@@ -1282,6 +1287,18 @@ their corresponding top-level category object in your `settings.json` file.
  - **Description:** Maximum number of directories to search for memory.
  - **Default:** `200`

+- **`context.memoryBoundaryMarkers`** (array):
+  - **Description:** File or directory names that mark the boundary for
+    GEMINI.md discovery. The upward traversal stops at the first directory
+    containing any of these markers. An empty array disables parent traversal.
+  - **Default:**
+
+    ```json
+    [".git"]
+    ```
+
+  - **Requires restart:** Yes
+
 - **`context.includeDirectories`** (array):
  - **Description:** Additional directories to include in the workspace context.
    Missing directories will be skipped with a warning.
@@ -1349,6 +1366,14 @@ their corresponding top-level category object in your `settings.json` file.
  - **Default:** `true`
  - **Requires restart:** Yes

+- **`tools.shell.backgroundCompletionBehavior`** (enum):
+  - **Description:** Controls what happens when a background shell command
+    finishes. 'silent' (default): quietly exits in background. 'inject':
+    automatically returns output to agent. 'notify': shows brief message in
+    chat.
+  - **Default:** `"silent"`
+  - **Values:** `"silent"`, `"inject"`, `"notify"`
+
 - **`tools.shell.pager`** (string):
  - **Description:** The pager command to use for shell output. Defaults to
    `cat`.
@@ -1677,6 +1702,28 @@ their corresponding top-level category object in your `settings.json` file.
  - **Default:** `false`
  - **Requires restart:** Yes

+- **`experimental.agentHistoryTruncation`** (boolean):
+  - **Description:** Enable truncation window logic for the Agent History
+    Provider.
+  - **Default:** `false`
+  - **Requires restart:** Yes
+
+- **`experimental.agentHistoryTruncationThreshold`** (number):
+  - **Description:** The maximum number of messages before history is truncated.
+  - **Default:** `30`
+  - **Requires restart:** Yes
+
+- **`experimental.agentHistoryRetainedMessages`** (number):
+  - **Description:** The number of recent messages to retain after truncation.
+  - **Default:** `15`
+  - **Requires restart:** Yes
+
+- **`experimental.agentHistorySummarization`** (boolean):
+  - **Description:** Enable summarization of truncated content via a small model
+    for the Agent History Provider.
+  - **Default:** `false`
+  - **Requires restart:** Yes
+
 - **`experimental.topicUpdateNarration`** (boolean):
  - **Description:** Enable the experimental Topic & Update communication model
    for reduced chattiness and structured progress reporting.
@@ -2160,37 +2207,14 @@ You can customize this behavior in your `settings.json` file:
 Arguments passed directly when running the CLI can override other configurations
 for that specific session.

- **`--model <model_name>`** (**`-m <model_name>`**):
-  - Specifies the Gemini model to use for this session.
-  - Example: `npm start -- --model gemini-3-pro-preview`
- **`--prompt <your_prompt>`** (**`-p <your_prompt>`**):
-  - **Deprecated:** Use positional arguments instead.
-  - Used to pass a prompt directly to the command. This invokes Gemini CLI in a
-    non-interactive mode.
- **`--prompt-interactive <your_prompt>`** (**`-i <your_prompt>`**):
-  - Starts an interactive session with the provided prompt as the initial input.
-  - The prompt is processed within the interactive session, not before it.
-  - Cannot be used when piping input from stdin.
-  - Example: `gemini -i "explain this code"`
- **`--output-format <format>`**:
-  - **Description:** Specifies the format of the CLI output for non-interactive
-    mode.
-  - **Values:**
-    - `text`: (Default) The standard human-readable output.
-    - `json`: A machine-readable JSON output.
-    - `stream-json`: A streaming JSON output that emits real-time events.
-  - **Note:** For structured output and scripting, use the
-    `--output-format json` or `--output-format stream-json` flag.
- **`--sandbox`** (**`-s`**):
-  - Enables sandbox mode for this session.
- **`--debug`** (**`-d`**):
-  - Enables debug mode for this session, providing more verbose output. Open the
-    debug console with F12 to see the additional logging.
-
- **`--help`** (or **`-h`**):
-  - Displays help information about command-line arguments.
- **`--yolo`**:
-  - Enables YOLO mode, which automatically approves all tool calls.
+- **`--acp`**:
+  - Starts the agent in Agent Communication Protocol (ACP) mode.
+- **`--allowed-mcp-server-names`**:
+  - A comma-separated list of MCP server names to allow for the session.
+- **`--allowed-tools <tool1,tool2,...>`**:
+  - A comma-separated list of tool names that will bypass the confirmation
+    dialog.
+  - Example: `gemini --allowed-tools "ShellTool(git status)"`
 - **`--approval-mode <mode>`**:
  - Sets the approval mode for tool calls. Available modes:
    - `default`: Prompt for approval on each tool call (default behavior)
@@ -2204,35 +2228,24 @@ for that specific session.
  - Cannot be used together with `--yolo`. Use `--approval-mode=yolo` instead of
    `--yolo` for the new unified approach.
  - Example: `gemini --approval-mode auto_edit`
- **`--allowed-tools <tool1,tool2,...>`**:
-  - A comma-separated list of tool names that will bypass the confirmation
-    dialog.
-  - Example: `gemini --allowed-tools "ShellTool(git status)"`
- **`--extensions <extension_name ...>`** (**`-e <extension_name ...>`**):
-  - Specifies a list of extensions to use for the session. If not provided, all
-    available extensions are used.
-  - Use the special term `gemini -e none` to disable all extensions.
-  - Example: `gemini -e my-extension -e my-other-extension`
- **`--list-extensions`** (**`-l`**):
-  - Lists all available extensions and exits.
- **`--resume [session_id]`** (**`-r [session_id]`**):
-  - Resume a previous chat session. Use "latest" for the most recent session,
-    provide a session index number, or provide a full session UUID.
-  - If no session_id is provided, defaults to "latest".
-  - Example: `gemini --resume 5` or `gemini --resume latest` or
-    `gemini --resume a1b2c3d4-e5f6-7890-abcd-ef1234567890` or `gemini --resume`
-  - See [Session Management](../cli/session-management.md) for more details.
- **`--list-sessions`**:
-  - List all available chat sessions for the current project and exit.
-  - Shows session indices, dates, message counts, and preview of first user
-    message.
-  - Example: `gemini --list-sessions`
+- **`--debug`** (**`-d`**):
+  - Enables debug mode for this session, providing more verbose output. Open the
+    debug console with F12 to see the additional logging.
 - **`--delete-session <identifier>`**:
  - Delete a specific chat session by its index number or full session UUID.
  - Use `--list-sessions` first to see available sessions, their indices, and
    UUIDs.
  - Example: `gemini --delete-session 3` or
    `gemini --delete-session a1b2c3d4-e5f6-7890-abcd-ef1234567890`
+- **`--extensions <extension_name ...>`** (**`-e <extension_name ...>`**):
+  - Specifies a list of extensions to use for the session. If not provided, all
+    available extensions are used.
+  - Use the special term `gemini -e none` to disable all extensions.
+  - Example: `gemini -e my-extension -e my-other-extension`
+- **`--fake-responses`**:
+  - Path to a file with fake model responses for testing.
+- **`--help`** (or **`-h`**):
+  - Displays help information about command-line arguments.
 - **`--include-directories <dir1,dir2,...>`**:
  - Includes additional directories in the workspace for multi-directory
    support.
@@ -2240,19 +2253,52 @@ for that specific session.
  - 5 directories can be added at maximum.
  - Example: `--include-directories /path/to/project1,/path/to/project2` or
    `--include-directories /path/to/project1 --include-directories /path/to/project2`
+- **`--list-extensions`** (**`-l`**):
+  - Lists all available extensions and exits.
+- **`--list-sessions`**:
+  - List all available chat sessions for the current project and exit.
+  - Shows session indices, dates, message counts, and preview of first user
+    message.
+  - Example: `gemini --list-sessions`
+- **`--model <model_name>`** (**`-m <model_name>`**):
+  - Specifies the Gemini model to use for this session.
+  - Example: `npm start -- --model gemini-3-pro-preview`
+- **`--output-format <format>`**:
+  - **Description:** Specifies the format of the CLI output for non-interactive
+    mode.
+  - **Values:**
+    - `text`: (Default) The standard human-readable output.
+    - `json`: A machine-readable JSON output.
+    - `stream-json`: A streaming JSON output that emits real-time events.
+  - **Note:** For structured output and scripting, use the
+    `--output-format json` or `--output-format stream-json` flag.
+- **`--prompt <your_prompt>`** (**`-p <your_prompt>`**):
+  - **Deprecated:** Use positional arguments instead.
+  - Used to pass a prompt directly to the command. This invokes Gemini CLI in a
+    non-interactive mode.
+- **`--prompt-interactive <your_prompt>`** (**`-i <your_prompt>`**):
+  - Starts an interactive session with the provided prompt as the initial input.
+  - The prompt is processed within the interactive session, not before it.
+  - Cannot be used when piping input from stdin.
+  - Example: `gemini -i "explain this code"`
+- **`--record-responses`**:
+  - Path to a file to record model responses for testing.
+- **`--resume [session_id]`** (**`-r [session_id]`**):
+  - Resume a previous chat session. Use "latest" for the most recent session,
+    provide a session index number, or provide a full session UUID.
+  - If no session_id is provided, defaults to "latest".
+  - Example: `gemini --resume 5` or `gemini --resume latest` or
+    `gemini --resume a1b2c3d4-e5f6-7890-abcd-ef1234567890` or `gemini --resume`
+  - See [Session Management](../cli/session-management.md) for more details.
+- **`--sandbox`** (**`-s`**):
+  - Enables sandbox mode for this session.
 - **`--screen-reader`**:
  - Enables screen reader mode, which adjusts the TUI for better compatibility
    with screen readers.
 - **`--version`**:
  - Displays the version of the CLI.
- **`--experimental-acp`**:
-  - Starts the agent in ACP mode.
- **`--allowed-mcp-server-names`**:
-  - Allowed MCP server names.
- **`--fake-responses`**:
-  - Path to a file with fake model responses for testing.
- **`--record-responses`**:
-  - Path to a file to record model responses for testing.
+- **`--yolo`**:
+  - Enables YOLO mode, which automatically approves all tool calls.

 ## Context files (hierarchical instructional context)

@@ -86,12 +86,13 @@ available combinations.

 #### Text Input

-| Command                    | Action                                                     | Keys                                                                                |
-| -------------------------- | ---------------------------------------------------------- | ----------------------------------------------------------------------------------- |
-| `input.submit`             | Submit the current prompt.                                 | `Enter`                                                                             |
-| `input.newline`            | Insert a newline without submitting.                       | `Ctrl+Enter`<br />`Cmd/Win+Enter`<br />`Alt+Enter`<br />`Shift+Enter`<br />`Ctrl+J` |
-| `input.openExternalEditor` | Open the current prompt or the plan in an external editor. | `Ctrl+X`                                                                            |
-| `input.paste`              | Paste from the clipboard.                                  | `Ctrl+V`<br />`Cmd/Win+V`<br />`Alt+V`                                              |
+| Command                    | Action                                                                    | Keys                                                                                |
+| -------------------------- | ------------------------------------------------------------------------- | ----------------------------------------------------------------------------------- |
+| `input.submit`             | Submit the current prompt.                                                | `Enter`                                                                             |
+| `input.queueMessage`       | Queue the current prompt to be processed after the current task finishes. | `Tab`                                                                               |
+| `input.newline`            | Insert a newline without submitting.                                      | `Ctrl+Enter`<br />`Cmd/Win+Enter`<br />`Alt+Enter`<br />`Shift+Enter`<br />`Ctrl+J` |
+| `input.openExternalEditor` | Open the current prompt or the plan in an external editor.                | `Ctrl+X`                                                                            |
+| `input.paste`              | Paste from the clipboard.                                                 | `Ctrl+V`<br />`Cmd/Win+V`<br />`Alt+V`                                              |

 #### App Controls

@@ -12,6 +12,21 @@ quota for your needs, see the [Plans page](https://geminicli.com/plans/).
 This article outlines the specific quotas and pricing applicable to Gemini CLI
 when using different authentication methods.

+The following table summarizes the available quotas and their respective limits:
+
+| Authentication method | Tier / Subscription             | Maximum requests per user per day |
+| :-------------------- | :------------------------------ | :-------------------------------- |
+| **Google account**    | Gemini Code Assist (Individual) | 1,000 requests                    |
+|                       | Google AI Pro                   | 1,500 requests                    |
+|                       | Google AI Ultra                 | 2,000 requests                    |
+| **Gemini API key**    | Free tier (Unpaid)              | 250 requests                      |
+|                       | Pay-as-you-go (Paid)            | Varies                            |
+| **Vertex AI**         | Express mode (Free)             | Varies                            |
+|                       | Pay-as-you-go (Paid)            | Varies                            |
+| **Google Workspace**  | Code Assist Standard            | 1,500 requests                    |
+|                       | Code Assist Enterprise          | 2,000 requests                    |
+|                       | Workspace AI Ultra              | 2,000 requests                    |
+
 Generally, there are three categories to choose from:

 - Free Usage: Ideal for experimentation and light use.
@@ -20,6 +35,9 @@ Generally, there are three categories to choose from:
 - Pay-As-You-Go: The most flexible option for professional use, long-running
  tasks, or when you need full control over your usage.

+Requests are limited per user per minute and are subject to the availability of
+the service in times of high demand.
+
 ## Free usage

 Access to Gemini CLI begins with a generous free tier, perfect for
@@ -33,8 +51,7 @@ authorization type.
 For users who authenticate by using their Google account to access Gemini Code
 Assist for individuals. This includes:

- 1000 model requests / user / day
- 60 model requests / user / minute
+- 1000 maximum model requests / user / day
 - Model requests will be made across the Gemini model family as determined by
  Gemini CLI.

@@ -46,8 +63,7 @@ Learn more at
 If you are using a Gemini API key, you can also benefit from a free tier. This
 includes:

- 250 model requests / user / day
- 10 model requests / user / minute
+- 250 maximum model requests / user / day
 - Model requests to Flash model only.

 Learn more at
@@ -59,7 +75,7 @@ Vertex AI offers an Express Mode without the need to enable billing. This
 includes:

 - 90 days before you need to enable billing.
- Quotas and models are variable and specific to your account.
+- Quotas and models are specific to your account and their limits vary.

 Learn more at
 [Vertex AI Express Mode Limits](https://cloud.google.com/vertex-ai/generative-ai/docs/start/express-mode/overview#quotas).
@@ -112,11 +128,9 @@ Standard/Plus and AI Expanded, are not supported._

  This includes the following request limits:
  - Gemini Code Assist Standard edition:
-    - 1500 model requests / user / day
-    - 120 model requests / user / minute
+    - 1500 maximum model requests / user / day
  - Gemini Code Assist Enterprise edition:
-    - 2000 model requests / user / day
-    - 120 model requests / user / minute
+    - 2000 maximum model requests / user / day
  - Model requests will be made across the Gemini model family as determined by
    Gemini CLI.

@@ -12,7 +12,6 @@
            "label": "Authentication",
            "slug": "docs/get-started/authentication"
          },
-          { "label": "Examples", "slug": "docs/get-started/examples" },
          { "label": "CLI cheatsheet", "slug": "docs/cli/cli-reference" },
          {
            "label": "Gemini 3 on Gemini CLI",
@@ -112,7 +111,17 @@
              { "label": "Reference", "slug": "docs/hooks/reference" }
            ]
          },
-          { "label": "IDE integration", "slug": "docs/ide-integration" },
+          {
+            "label": "IDE integration",
+            "collapsed": true,
+            "items": [
+              { "label": "Overview", "slug": "docs/ide-integration" },
+              {
+                "label": "Developer guide: ACP mode",
+                "slug": "docs/cli/acp-mode"
+              }
+            ]
+          },
          { "label": "MCP servers", "slug": "docs/tools/mcp-server" },
          { "label": "Model routing", "slug": "docs/cli/model-routing" },
          { "label": "Model selection", "slug": "docs/cli/model" },
@@ -13,8 +13,21 @@ import { evalTest, TEST_AGENTS } from './test-helper.js';

 const INDEX_TS = 'export const add = (a: number, b: number) => a + b;\n';

+// A minimal package.json is used to provide a realistic workspace anchor.
+// This prevents the agent from making incorrect assumptions about the environment
+// and helps it properly navigate or act as if it is in a standard Node.js project.
+const MOCK_PACKAGE_JSON = JSON.stringify(
+  {
+    name: 'subagent-eval-project',
+    version: '1.0.0',
+    type: 'module',
+  },
+  null,
+  2,
+);
+
 function readProjectFile(
-  rig: { testDir?: string },
+  rig: { testDir: string | null },
  relativePath: string,
 ): string {
  return fs.readFileSync(path.join(rig.testDir!, relativePath), 'utf8');
@@ -117,15 +130,7 @@ describe('subagent eval test cases', () => {
    files: {
      ...TEST_AGENTS.TESTING_AGENT.asFile(),
      'index.ts': INDEX_TS,
-      'package.json': JSON.stringify(
-        {
-          name: 'subagent-eval-project',
-          version: '1.0.0',
-          type: 'module',
-        },
-        null,
-        2,
-      ),
+      'package.json': MOCK_PACKAGE_JSON,
    },
    assert: async (rig, _result) => {
      const toolLogs = rig.readToolLogs() as Array<{
@@ -164,15 +169,7 @@ describe('subagent eval test cases', () => {
      ...TEST_AGENTS.TESTING_AGENT.asFile(),
      'index.ts': INDEX_TS,
      'README.md': 'TODO: update the README.\n',
-      'package.json': JSON.stringify(
-        {
-          name: 'subagent-eval-project',
-          version: '1.0.0',
-          type: 'module',
-        },
-        null,
-        2,
-      ),
+      'package.json': MOCK_PACKAGE_JSON,
    },
    assert: async (rig, _result) => {
      const toolLogs = rig.readToolLogs() as Array<{
@@ -190,4 +187,105 @@ describe('subagent eval test cases', () => {
      );
    },
  });
+
+  /**
+   * Checks that the main agent can correctly select the appropriate subagent
+   * from a large pool of available subagents (10 total).
+   */
+  evalTest('USUALLY_PASSES', {
+    name: 'should select the correct subagent from a pool of 10 different agents',
+    prompt: 'Please add a new SQL table migration for a user profile.',
+    files: {
+      ...TEST_AGENTS.DOCS_AGENT.asFile(),
+      ...TEST_AGENTS.TESTING_AGENT.asFile(),
+      ...TEST_AGENTS.DATABASE_AGENT.asFile(),
+      ...TEST_AGENTS.CSS_AGENT.asFile(),
+      ...TEST_AGENTS.I18N_AGENT.asFile(),
+      ...TEST_AGENTS.SECURITY_AGENT.asFile(),
+      ...TEST_AGENTS.DEVOPS_AGENT.asFile(),
+      ...TEST_AGENTS.ANALYTICS_AGENT.asFile(),
+      ...TEST_AGENTS.ACCESSIBILITY_AGENT.asFile(),
+      ...TEST_AGENTS.MOBILE_AGENT.asFile(),
+      'package.json': MOCK_PACKAGE_JSON,
+    },
+    assert: async (rig, _result) => {
+      const toolLogs = rig.readToolLogs() as Array<{
+        toolRequest: { name: string };
+      }>;
+      await rig.expectToolCallSuccess(['database-agent']);
+
+      // Ensure the generalist and other irrelevant specialists were not invoked
+      const uncalledAgents = [
+        'generalist',
+        TEST_AGENTS.DOCS_AGENT.name,
+        TEST_AGENTS.TESTING_AGENT.name,
+        TEST_AGENTS.CSS_AGENT.name,
+        TEST_AGENTS.I18N_AGENT.name,
+        TEST_AGENTS.SECURITY_AGENT.name,
+        TEST_AGENTS.DEVOPS_AGENT.name,
+        TEST_AGENTS.ANALYTICS_AGENT.name,
+        TEST_AGENTS.ACCESSIBILITY_AGENT.name,
+        TEST_AGENTS.MOBILE_AGENT.name,
+      ];
+
+      for (const agentName of uncalledAgents) {
+        expect(toolLogs.some((l) => l.toolRequest.name === agentName)).toBe(
+          false,
+        );
+      }
+    },
+  });
+
+  /**
+   * Checks that the main agent can correctly select the appropriate subagent
+   * from a large pool of available subagents, even when many irrelevant MCP tools are present.
+   *
+   * This test includes stress tests the subagent delegation with ~80 tools.
+   */
+  evalTest('USUALLY_PASSES', {
+    name: 'should select the correct subagent from a pool of 10 different agents with MCP tools present',
+    prompt: 'Please add a new SQL table migration for a user profile.',
+    setup: async (rig) => {
+      rig.addTestMcpServer('workspace-server', 'google-workspace');
+    },
+    files: {
+      ...TEST_AGENTS.DOCS_AGENT.asFile(),
+      ...TEST_AGENTS.TESTING_AGENT.asFile(),
+      ...TEST_AGENTS.DATABASE_AGENT.asFile(),
+      ...TEST_AGENTS.CSS_AGENT.asFile(),
+      ...TEST_AGENTS.I18N_AGENT.asFile(),
+      ...TEST_AGENTS.SECURITY_AGENT.asFile(),
+      ...TEST_AGENTS.DEVOPS_AGENT.asFile(),
+      ...TEST_AGENTS.ANALYTICS_AGENT.asFile(),
+      ...TEST_AGENTS.ACCESSIBILITY_AGENT.asFile(),
+      ...TEST_AGENTS.MOBILE_AGENT.asFile(),
+      'package.json': MOCK_PACKAGE_JSON,
+    },
+    assert: async (rig, _result) => {
+      const toolLogs = rig.readToolLogs() as Array<{
+        toolRequest: { name: string };
+      }>;
+      await rig.expectToolCallSuccess(['database-agent']);
+
+      // Ensure the generalist and other irrelevant specialists were not invoked
+      const uncalledAgents = [
+        'generalist',
+        TEST_AGENTS.DOCS_AGENT.name,
+        TEST_AGENTS.TESTING_AGENT.name,
+        TEST_AGENTS.CSS_AGENT.name,
+        TEST_AGENTS.I18N_AGENT.name,
+        TEST_AGENTS.SECURITY_AGENT.name,
+        TEST_AGENTS.DEVOPS_AGENT.name,
+        TEST_AGENTS.ANALYTICS_AGENT.name,
+        TEST_AGENTS.ACCESSIBILITY_AGENT.name,
+        TEST_AGENTS.MOBILE_AGENT.name,
+      ];
+
+      for (const agentName of uncalledAgents) {
+        expect(toolLogs.some((l) => l.toolRequest.name === agentName)).toBe(
+          false,
+        );
+      }
+    },
+  });
 });
@@ -0,0 +1,207 @@
+/**
+ * @license
+ * Copyright 2026 Google LLC
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+import { describe, it, expect, vi, beforeEach, afterEach } from 'vitest';
+import fs from 'node:fs';
+import path from 'node:path';
+import { internalEvalTest } from './test-helper.js';
+import { TestRig } from '@google/gemini-cli-test-utils';
+
+// Mock TestRig to control API success/failure
+vi.mock('@google/gemini-cli-test-utils', () => {
+  return {
+    TestRig: vi.fn().mockImplementation(() => ({
+      setup: vi.fn(),
+      run: vi.fn(),
+      cleanup: vi.fn(),
+      readToolLogs: vi.fn().mockReturnValue([]),
+      _lastRunStderr: '',
+    })),
+  };
+});
+
+describe('evalTest reliability logic', () => {
+  const LOG_DIR = path.resolve(process.cwd(), 'evals/logs');
+  const RELIABILITY_LOG = path.join(LOG_DIR, 'api-reliability.jsonl');
+
+  beforeEach(() => {
+    vi.clearAllMocks();
+    if (fs.existsSync(RELIABILITY_LOG)) {
+      fs.unlinkSync(RELIABILITY_LOG);
+    }
+  });
+
+  afterEach(() => {
+    if (fs.existsSync(RELIABILITY_LOG)) {
+      fs.unlinkSync(RELIABILITY_LOG);
+    }
+  });
+
+  it('should retry 3 times on 500 INTERNAL error and then SKIP', async () => {
+    const mockRig = new TestRig() as any;
+    (TestRig as any).mockReturnValue(mockRig);
+
+    // Simulate permanent 500 error
+    mockRig.run.mockRejectedValue(new Error('status: INTERNAL - API Down'));
+
+    // Execute the test function directly
+    await internalEvalTest({
+      name: 'test-api-failure',
+      prompt: 'do something',
+      assert: async () => {},
+    });
+
+    // Verify retries: 1 initial + 3 retries = 4 setups/runs
+    expect(mockRig.run).toHaveBeenCalledTimes(4);
+
+    // Verify log content
+    const logContent = fs
+      .readFileSync(RELIABILITY_LOG, 'utf-8')
+      .trim()
+      .split('\n');
+    expect(logContent.length).toBe(4);
+
+    const entries = logContent.map((line) => JSON.parse(line));
+    expect(entries[0].status).toBe('RETRY');
+    expect(entries[0].attempt).toBe(0);
+    expect(entries[3].status).toBe('SKIP');
+    expect(entries[3].attempt).toBe(3);
+    expect(entries[3].testName).toBe('test-api-failure');
+  });
+
+  it('should fail immediately on non-500 errors (like assertion failures)', async () => {
+    const mockRig = new TestRig() as any;
+    (TestRig as any).mockReturnValue(mockRig);
+
+    // Simulate a real logic error/bug
+    mockRig.run.mockResolvedValue('Success');
+    const assertError = new Error('Assertion failed: expected foo to be bar');
+
+    // Expect the test function to throw immediately
+    await expect(
+      internalEvalTest({
+        name: 'test-logic-failure',
+        prompt: 'do something',
+        assert: async () => {
+          throw assertError;
+        },
+      }),
+    ).rejects.toThrow('Assertion failed');
+
+    // Verify NO retries: only 1 attempt
+    expect(mockRig.run).toHaveBeenCalledTimes(1);
+
+    // Verify NO reliability log was created (it's not an API error)
+    expect(fs.existsSync(RELIABILITY_LOG)).toBe(false);
+  });
+
+  it('should recover if a retry succeeds', async () => {
+    const mockRig = new TestRig() as any;
+    (TestRig as any).mockReturnValue(mockRig);
+
+    // Fail once, then succeed
+    mockRig.run
+      .mockRejectedValueOnce(new Error('status: INTERNAL'))
+      .mockResolvedValueOnce('Success');
+
+    await internalEvalTest({
+      name: 'test-recovery',
+      prompt: 'do something',
+      assert: async () => {},
+    });
+
+    // Ran twice: initial (fail) + retry 1 (success)
+    expect(mockRig.run).toHaveBeenCalledTimes(2);
+
+    // Log should only have the one RETRY entry
+    const logContent = fs
+      .readFileSync(RELIABILITY_LOG, 'utf-8')
+      .trim()
+      .split('\n');
+    expect(logContent.length).toBe(1);
+    expect(JSON.parse(logContent[0]).status).toBe('RETRY');
+  });
+
+  it('should retry 3 times on 503 UNAVAILABLE error and then SKIP', async () => {
+    const mockRig = new TestRig() as any;
+    (TestRig as any).mockReturnValue(mockRig);
+
+    // Simulate permanent 503 error
+    mockRig.run.mockRejectedValue(
+      new Error('status: UNAVAILABLE - Service Busy'),
+    );
+
+    await internalEvalTest({
+      name: 'test-api-503',
+      prompt: 'do something',
+      assert: async () => {},
+    });
+
+    expect(mockRig.run).toHaveBeenCalledTimes(4);
+
+    const logContent = fs
+      .readFileSync(RELIABILITY_LOG, 'utf-8')
+      .trim()
+      .split('\n');
+    const entries = logContent.map((line) => JSON.parse(line));
+    expect(entries[0].errorCode).toBe('503');
+    expect(entries[3].status).toBe('SKIP');
+  });
+
+  it('should throw if an absolute path is used in files', async () => {
+    const mockRig = new TestRig() as any;
+    (TestRig as any).mockReturnValue(mockRig);
+    mockRig.testDir = path.resolve(process.cwd(), 'test-dir-tmp');
+    if (!fs.existsSync(mockRig.testDir)) {
+      fs.mkdirSync(mockRig.testDir, { recursive: true });
+    }
+
+    try {
+      await expect(
+        internalEvalTest({
+          name: 'test-absolute-path',
+          prompt: 'do something',
+          files: {
+            '/etc/passwd': 'hacked',
+          },
+          assert: async () => {},
+        }),
+      ).rejects.toThrow('Invalid file path in test case: /etc/passwd');
+    } finally {
+      if (fs.existsSync(mockRig.testDir)) {
+        fs.rmSync(mockRig.testDir, { recursive: true, force: true });
+      }
+    }
+  });
+
+  it('should throw if directory traversal is detected in files', async () => {
+    const mockRig = new TestRig() as any;
+    (TestRig as any).mockReturnValue(mockRig);
+    mockRig.testDir = path.resolve(process.cwd(), 'test-dir-tmp');
+
+    // Create a mock test-dir
+    if (!fs.existsSync(mockRig.testDir)) {
+      fs.mkdirSync(mockRig.testDir, { recursive: true });
+    }
+
+    try {
+      await expect(
+        internalEvalTest({
+          name: 'test-traversal',
+          prompt: 'do something',
+          files: {
+            '../sensitive.txt': 'hacked',
+          },
+          assert: async () => {},
+        }),
+      ).rejects.toThrow('Invalid file path in test case: ../sensitive.txt');
+    } finally {
+      if (fs.existsSync(mockRig.testDir)) {
+        fs.rmSync(mockRig.testDir, { recursive: true, force: true });
+      }
+    }
+  });
+});
@@ -39,87 +39,38 @@ export * from '@google/gemini-cli-test-utils';
 export type EvalPolicy = 'ALWAYS_PASSES' | 'USUALLY_PASSES';

 export function evalTest(policy: EvalPolicy, evalCase: EvalCase) {
-  const fn = async () => {
+  runEval(
+    policy,
+    evalCase.name,
+    () => internalEvalTest(evalCase),
+    evalCase.timeout,
+  );
+}
+
+export async function internalEvalTest(evalCase: EvalCase) {
+  const maxRetries = 3;
+  let attempt = 0;
+
+  while (attempt <= maxRetries) {
    const rig = new TestRig();
    const { logDir, sanitizedName } = await prepareLogDir(evalCase.name);
    const activityLogFile = path.join(logDir, `${sanitizedName}.jsonl`);
    const logFile = path.join(logDir, `${sanitizedName}.log`);
    let isSuccess = false;
+
    try {
      rig.setup(evalCase.name, evalCase.params);

-      // Symlink node modules to reduce the amount of time needed to
-      // bootstrap test projects.
-      symlinkNodeModules(rig.testDir || '');
+      if (evalCase.setup) {
+        await evalCase.setup(rig);
+      }

      if (evalCase.files) {
-        const acknowledgedAgents: Record<string, Record<string, string>> = {};
-        const projectRoot = fs.realpathSync(rig.testDir!);
-
-        for (const [filePath, content] of Object.entries(evalCase.files)) {
-          const fullPath = path.join(rig.testDir!, filePath);
-          fs.mkdirSync(path.dirname(fullPath), { recursive: true });
-          fs.writeFileSync(fullPath, content);
-
-          // If it's an agent file, calculate hash for acknowledgement
-          if (
-            filePath.startsWith('.gemini/agents/') &&
-            filePath.endsWith('.md')
-          ) {
-            const hash = crypto
-              .createHash('sha256')
-              .update(content)
-              .digest('hex');
-
-            try {
-              const agentDefs = await parseAgentMarkdown(fullPath, content);
-              if (agentDefs.length > 0) {
-                const agentName = agentDefs[0].name;
-                if (!acknowledgedAgents[projectRoot]) {
-                  acknowledgedAgents[projectRoot] = {};
-                }
-                acknowledgedAgents[projectRoot][agentName] = hash;
-              }
-            } catch (error) {
-              console.warn(
-                `Failed to parse agent for test acknowledgement: ${filePath}`,
-                error,
-              );
-            }
-          }
-        }
-
-        // Write acknowledged_agents.json to the home directory
-        if (Object.keys(acknowledgedAgents).length > 0) {
-          const ackPath = path.join(
-            rig.homeDir!,
-            '.gemini',
-            'acknowledgments',
-            'agents.json',
-          );
-          fs.mkdirSync(path.dirname(ackPath), { recursive: true });
-          fs.writeFileSync(
-            ackPath,
-            JSON.stringify(acknowledgedAgents, null, 2),
-          );
-        }
-
-        const execOptions = { cwd: rig.testDir!, stdio: 'inherit' as const };
-        execSync('git init', execOptions);
-        execSync('git config user.email "test@example.com"', execOptions);
-        execSync('git config user.name "Test User"', execOptions);
-
-        // Temporarily disable the interactive editor and git pager
-        // to avoid hanging the tests. It seems the the agent isn't
-        // consistently honoring the instructions to avoid interactive
-        // commands.
-        execSync('git config core.editor "true"', execOptions);
-        execSync('git config core.pager "cat"', execOptions);
-        execSync('git config commit.gpgsign false', execOptions);
-        execSync('git add .', execOptions);
-        execSync('git commit --allow-empty -m "Initial commit"', execOptions);
+        await setupTestFiles(rig, evalCase.files);
      }

+      symlinkNodeModules(rig.testDir || '');
+
      // If messages are provided, write a session file so --resume can load it.
      let sessionId: string | undefined;
      if (evalCase.messages) {
@@ -188,6 +139,37 @@ export function evalTest(policy: EvalPolicy, evalCase: EvalCase) {

      await evalCase.assert(rig, result);
      isSuccess = true;
+      return; // Success! Exit the retry loop.
+    } catch (error: unknown) {
+      const errorMessage =
+        error instanceof Error ? error.message : String(error);
+      const errorCode = getApiErrorCode(errorMessage);
+
+      if (errorCode) {
+        const status = attempt < maxRetries ? 'RETRY' : 'SKIP';
+        logReliabilityEvent(
+          evalCase.name,
+          attempt,
+          status,
+          errorCode,
+          errorMessage,
+        );
+
+        if (attempt < maxRetries) {
+          attempt++;
+          console.warn(
+            `[Eval] Attempt ${attempt} failed with ${errorCode} Error. Retrying...`,
+          );
+          continue; // Retry
+        }
+
+        console.warn(
+          `[Eval] '${evalCase.name}' failed after ${maxRetries} retries due to persistent API errors. Skipping failure to avoid blocking PR.`,
+        );
+        return; // Gracefully exit without failing the test
+      }
+
+      throw error; // Real failure
    } finally {
      if (isSuccess) {
        await fs.promises.unlink(activityLogFile).catch((err) => {
@@ -206,9 +188,131 @@ export function evalTest(policy: EvalPolicy, evalCase: EvalCase) {
      );
      await rig.cleanup();
    }
+  }
+}
+
+function getApiErrorCode(message: string): '500' | '503' | undefined {
+  if (
+    message.includes('status: UNAVAILABLE') ||
+    message.includes('code: 503') ||
+    message.includes('Service Unavailable')
+  ) {
+    return '503';
+  }
+  if (
+    message.includes('status: INTERNAL') ||
+    message.includes('code: 500') ||
+    message.includes('Internal error encountered')
+  ) {
+    return '500';
+  }
+  return undefined;
+}
+
+/**
+ * Log reliability event for later harvesting.
+ *
+ * Note: Uses synchronous file I/O to ensure the log is persisted even if the
+ * test process is abruptly terminated by a timeout or CI crash. Performance
+ * impact is negligible compared to long-running evaluation tests.
+ */
+function logReliabilityEvent(
+  testName: string,
+  attempt: number,
+  status: 'RETRY' | 'SKIP',
+  errorCode: '500' | '503',
+  errorMessage: string,
+) {
+  const reliabilityLog = {
+    timestamp: new Date().toISOString(),
+    testName,
+    model: process.env.GEMINI_MODEL || 'unknown',
+    attempt,
+    status,
+    errorCode,
+    error: errorMessage,
  };

-  runEval(policy, evalCase.name, fn, evalCase.timeout);
+  try {
+    const relDir = path.resolve(process.cwd(), 'evals/logs');
+    fs.mkdirSync(relDir, { recursive: true });
+    fs.appendFileSync(
+      path.join(relDir, 'api-reliability.jsonl'),
+      JSON.stringify(reliabilityLog) + '\n',
+    );
+  } catch (logError) {
+    console.error('Failed to write reliability log:', logError);
+  }
+}
+
+/**
+ * Helper to setup test files and git repository.
+ *
+ * Note: While this is an async function (due to parseAgentMarkdown), it
+ * intentionally uses synchronous filesystem and child_process operations
+ * for simplicity and to ensure sequential environment preparation.
+ */
+async function setupTestFiles(rig: TestRig, files: Record<string, string>) {
+  const acknowledgedAgents: Record<string, Record<string, string>> = {};
+  const projectRoot = fs.realpathSync(rig.testDir!);
+
+  for (const [filePath, content] of Object.entries(files)) {
+    if (filePath.includes('..') || path.isAbsolute(filePath)) {
+      throw new Error(`Invalid file path in test case: ${filePath}`);
+    }
+    const fullPath = path.join(projectRoot, filePath);
+    if (!fullPath.startsWith(projectRoot)) {
+      throw new Error(`Path traversal detected: ${filePath}`);
+    }
+
+    fs.mkdirSync(path.dirname(fullPath), { recursive: true });
+    fs.writeFileSync(fullPath, content);
+
+    if (filePath.startsWith('.gemini/agents/') && filePath.endsWith('.md')) {
+      const hash = crypto.createHash('sha256').update(content).digest('hex');
+      try {
+        const agentDefs = await parseAgentMarkdown(fullPath, content);
+        if (agentDefs.length > 0) {
+          const agentName = agentDefs[0].name;
+          if (!acknowledgedAgents[projectRoot]) {
+            acknowledgedAgents[projectRoot] = {};
+          }
+          acknowledgedAgents[projectRoot][agentName] = hash;
+        }
+      } catch (error) {
+        console.warn(
+          `Failed to parse agent for test acknowledgement: ${filePath}`,
+          error,
+        );
+      }
+    }
+  }
+
+  if (Object.keys(acknowledgedAgents).length > 0) {
+    const ackPath = path.join(
+      rig.homeDir!,
+      '.gemini',
+      'acknowledgments',
+      'agents.json',
+    );
+    fs.mkdirSync(path.dirname(ackPath), { recursive: true });
+    fs.writeFileSync(ackPath, JSON.stringify(acknowledgedAgents, null, 2));
+  }
+
+  const execOptions = { cwd: rig.testDir!, stdio: 'inherit' as const };
+  execSync('git init --initial-branch=main', execOptions);
+  execSync('git config user.email "test@example.com"', execOptions);
+  execSync('git config user.name "Test User"', execOptions);
+
+  // Temporarily disable the interactive editor and git pager
+  // to avoid hanging the tests. It seems the the agent isn't
+  // consistently honoring the instructions to avoid interactive
+  // commands.
+  execSync('git config core.editor "true"', execOptions);
+  execSync('git config core.pager "cat"', execOptions);
+  execSync('git config commit.gpgsign false', execOptions);
+  execSync('git add .', execOptions);
+  execSync('git commit --allow-empty -m "Initial commit"', execOptions);
 }

 /**
@@ -271,6 +375,7 @@ export interface EvalCase {
  prompt: string;
  timeout?: number;
  files?: Record<string, string>;
+  setup?: (rig: TestRig) => Promise<void> | void;
  /** Conversation history to pre-load via --resume. Each entry is a message object with type, content, etc. */
  messages?: Record<string, unknown>[];
  /** Session ID for the resumed session. Auto-generated if not provided. */
@@ -16,10 +16,6 @@ export default defineConfig({
  },
  test: {
    testTimeout: 300000, // 5 minutes
-    // Retry in CI but not nightly to avoid blocking on API error.
-    retry: process.env['VITEST_RETRY']
-      ? parseInt(process.env['VITEST_RETRY'], 10)
-      : 3,
    reporters: ['default', 'json'],
    outputFile: {
      json: 'evals/logs/report.json',
@@ -10,13 +10,9 @@ import { TestMcpServer } from './test-mcp-server.js';
 import { writeFileSync } from 'node:fs';
 import { join } from 'node:path';
 import { safeJsonStringify } from '@google/gemini-cli-core/src/utils/safeJsonStringify.js';
-import { env } from 'node:process';
-import { platform } from 'node:os';

 import stripAnsi from 'strip-ansi';

-const itIf = (condition: boolean) => (condition ? it : it.skip);
-
 describe('extension reloading', () => {
  let rig: TestRig;

@@ -26,141 +22,130 @@ describe('extension reloading', () => {

  afterEach(async () => await rig.cleanup());

-  const sandboxEnv = env['GEMINI_SANDBOX'];
-  // Fails in linux non-sandbox e2e tests
+  // always fails
  // TODO(#14527): Re-enable this once fixed
-  // Fails in sandbox mode, can't check for local extension updates.
-  itIf(
-    (!sandboxEnv || sandboxEnv === 'false') &&
-      platform() !== 'win32' &&
-      platform() !== 'linux',
-  )(
-    'installs a local extension, updates it, checks it was reloaded properly',
-    async () => {
-      const serverA = new TestMcpServer();
-      const portA = await serverA.start({
-        hello: () => ({ content: [{ type: 'text', text: 'world' }] }),
-      });
-      const extension = {
-        name: 'test-extension',
-        version: '0.0.1',
-        mcpServers: {
-          'test-server': {
-            httpUrl: `http://localhost:${portA}/mcp`,
-          },
+  it.skip('installs a local extension, updates it, checks it was reloaded properly', async () => {
+    const serverA = new TestMcpServer();
+    const portA = await serverA.start({
+      hello: () => ({ content: [{ type: 'text', text: 'world' }] }),
+    });
+    const extension = {
+      name: 'test-extension',
+      version: '0.0.1',
+      mcpServers: {
+        'test-server': {
+          httpUrl: `http://localhost:${portA}/mcp`,
        },
-      };
+      },
+    };

-      rig.setup('extension reload test', {
-        settings: {
-          experimental: { extensionReloading: true },
-        },
-      });
-      const testServerPath = join(rig.testDir!, 'gemini-extension.json');
-      writeFileSync(testServerPath, safeJsonStringify(extension, 2));
-      // defensive cleanup from previous tests.
-      try {
-        await rig.runCommand(['extensions', 'uninstall', 'test-extension']);
-      } catch {
-        /* empty */
-      }
-
-      const result = await rig.runCommand(
-        ['--debug', 'extensions', 'install', `${rig.testDir!}`],
-        { stdin: 'y\n' },
-      );
-      expect(result).toContain('test-extension');
-
-      // Now create the update, but its not installed yet
-      const serverB = new TestMcpServer();
-      const portB = await serverB.start({
-        goodbye: () => ({ content: [{ type: 'text', text: 'world' }] }),
-      });
-      extension.version = '0.0.2';
-      extension.mcpServers['test-server'].httpUrl =
-        `http://localhost:${portB}/mcp`;
-      writeFileSync(testServerPath, safeJsonStringify(extension, 2));
-
-      // Start the CLI.
-      const run = await rig.runInteractive({ args: '--debug' });
-      await run.expectText('You have 1 extension with an update available');
-      // See the outdated extension
-      await run.sendText('/extensions list');
-      await run.type('\r');
-      await run.expectText(
-        'test-extension (v0.0.1) - active (update available)',
-      );
-      // Wait for the UI to settle and retry the command until we see the update
-      await new Promise((resolve) => setTimeout(resolve, 1000));
-
-      // Poll for the updated list
-      await rig.pollCommand(
-        async () => {
-          await run.sendText('/mcp list');
-          await run.type('\r');
-        },
-        () => {
-          const output = stripAnsi(run.output);
-          return (
-            output.includes(
-              'test-server (from test-extension) - Ready (1 tool)',
-            ) && output.includes('- mcp_test-server_hello')
-          );
-        },
-        30000, // 30s timeout
-      );
-
-      // Update the extension, expect the list to update, and mcp servers as well.
-      await run.sendKeys('\u0015/extensions update test-extension');
-      await run.expectText('/extensions update test-extension');
-      await run.type('\r');
-      await new Promise((resolve) => setTimeout(resolve, 500));
-      await run.type('\r');
-      await run.expectText(
-        ` * test-server (remote): http://localhost:${portB}/mcp`,
-      );
-      await run.type('\r'); // consent
-      await run.expectText(
-        'Extension "test-extension" successfully updated: 0.0.1 → 0.0.2',
-      );
-
-      // Poll for the updated extension version
-      await rig.pollCommand(
-        async () => {
-          await run.sendText('/extensions list');
-          await run.type('\r');
-        },
-        () =>
-          stripAnsi(run.output).includes(
-            'test-extension (v0.0.2) - active (updated)',
-          ),
-        30000,
-      );
-
-      // Poll for the updated mcp tool
-      await rig.pollCommand(
-        async () => {
-          await run.sendText('/mcp list');
-          await run.type('\r');
-        },
-        () => {
-          const output = stripAnsi(run.output);
-          return (
-            output.includes(
-              'test-server (from test-extension) - Ready (1 tool)',
-            ) && output.includes('- mcp_test-server_goodbye')
-          );
-        },
-        30000,
-      );
-
-      await run.sendText('/quit');
-      await run.type('\r');
-
-      // Clean things up.
-      await serverA.stop();
-      await serverB.stop();
+    rig.setup('extension reload test', {
+      settings: {
+        experimental: { extensionReloading: true },
+      },
+    });
+    const testServerPath = join(rig.testDir!, 'gemini-extension.json');
+    writeFileSync(testServerPath, safeJsonStringify(extension, 2));
+    // defensive cleanup from previous tests.
+    try {
      await rig.runCommand(['extensions', 'uninstall', 'test-extension']);
-    },
-  );
+    } catch {
+      /* empty */
+    }
+
+    const result = await rig.runCommand(
+      ['--debug', 'extensions', 'install', `${rig.testDir!}`],
+      { stdin: 'y\n' },
+    );
+    expect(result).toContain('test-extension');
+
+    // Now create the update, but its not installed yet
+    const serverB = new TestMcpServer();
+    const portB = await serverB.start({
+      goodbye: () => ({ content: [{ type: 'text', text: 'world' }] }),
+    });
+    extension.version = '0.0.2';
+    extension.mcpServers['test-server'].httpUrl =
+      `http://localhost:${portB}/mcp`;
+    writeFileSync(testServerPath, safeJsonStringify(extension, 2));
+
+    // Start the CLI.
+    const run = await rig.runInteractive({ args: '--debug' });
+    await run.expectText('You have 1 extension with an update available');
+    // See the outdated extension
+    await run.sendText('/extensions list');
+    await run.type('\r');
+    await run.expectText('test-extension (v0.0.1) - active (update available)');
+    // Wait for the UI to settle and retry the command until we see the update
+    await new Promise((resolve) => setTimeout(resolve, 1000));
+
+    // Poll for the updated list
+    await rig.pollCommand(
+      async () => {
+        await run.sendText('/mcp list');
+        await run.type('\r');
+      },
+      () => {
+        const output = stripAnsi(run.output);
+        return (
+          output.includes(
+            'test-server (from test-extension) - Ready (1 tool)',
+          ) && output.includes('- mcp_test-server_hello')
+        );
+      },
+      30000, // 30s timeout
+    );
+
+    // Update the extension, expect the list to update, and mcp servers as well.
+    await run.sendKeys('\u0015/extensions update test-extension');
+    await run.expectText('/extensions update test-extension');
+    await run.type('\r');
+    await new Promise((resolve) => setTimeout(resolve, 500));
+    await run.type('\r');
+    await run.expectText(
+      ` * test-server (remote): http://localhost:${portB}/mcp`,
+    );
+    await run.type('\r'); // consent
+    await run.expectText(
+      'Extension "test-extension" successfully updated: 0.0.1 → 0.0.2',
+    );
+
+    // Poll for the updated extension version
+    await rig.pollCommand(
+      async () => {
+        await run.sendText('/extensions list');
+        await run.type('\r');
+      },
+      () =>
+        stripAnsi(run.output).includes(
+          'test-extension (v0.0.2) - active (updated)',
+        ),
+      30000,
+    );
+
+    // Poll for the updated mcp tool
+    await rig.pollCommand(
+      async () => {
+        await run.sendText('/mcp list');
+        await run.type('\r');
+      },
+      () => {
+        const output = stripAnsi(run.output);
+        return (
+          output.includes(
+            'test-server (from test-extension) - Ready (1 tool)',
+          ) && output.includes('- mcp_test-server_goodbye')
+        );
+      },
+      30000,
+    );
+
+    await run.sendText('/quit');
+    await run.type('\r');
+
+    // Clean things up.
+    await serverA.stop();
+    await serverB.stop();
+    await rig.runCommand(['extensions', 'uninstall', 'test-extension']);
+  });
 });
@@ -4,8 +4,10 @@
 * SPDX-License-Identifier: Apache-2.0
 */

+import { writeFileSync, mkdirSync } from 'node:fs';
+import { join } from 'node:path';
 import { describe, it, expect, beforeEach, afterEach } from 'vitest';
-import { TestRig, checkModelOutputContent } from './test-helper.js';
+import { GEMINI_DIR, TestRig, checkModelOutputContent } from './test-helper.js';

 describe('Plan Mode', () => {
  let rig: TestRig;
@@ -227,4 +229,68 @@ describe('Plan Mode', () => {
      `Expected write_file to succeed, but it failed with error: ${planWrite?.toolRequest.error}`,
    ).toBe(true);
  });
+  it('should switch from a pro model to a flash model after exiting plan mode', async () => {
+    const plansDir = 'plans-folder';
+    const planFilename = 'my-plan.md';
+
+    await rig.setup('should-switch-to-flash', {
+      settings: {
+        model: {
+          name: 'auto-gemini-2.5',
+        },
+        experimental: { plan: true },
+        tools: {
+          core: ['exit_plan_mode', 'run_shell_command'],
+          allowed: ['exit_plan_mode', 'run_shell_command'],
+        },
+        general: {
+          defaultApprovalMode: 'plan',
+          plan: {
+            directory: plansDir,
+          },
+        },
+      },
+    });
+
+    writeFileSync(
+      join(rig.homeDir!, GEMINI_DIR, 'state.json'),
+      JSON.stringify({ terminalSetupPromptShown: true }, null, 2),
+    );
+
+    const fullPlansDir = join(rig.testDir!, plansDir);
+    mkdirSync(fullPlansDir, { recursive: true });
+    writeFileSync(join(fullPlansDir, planFilename), 'Execute echo hello');
+
+    await rig.run({
+      approvalMode: 'plan',
+      stdin: `Exit plan mode using ${planFilename} and then run a shell command \`echo hello\`.`,
+    });
+
+    const exitCallFound = await rig.waitForToolCall('exit_plan_mode');
+    expect(exitCallFound, 'Expected exit_plan_mode to be called').toBe(true);
+
+    const shellCallFound = await rig.waitForToolCall('run_shell_command');
+    expect(shellCallFound, 'Expected run_shell_command to be called').toBe(
+      true,
+    );
+
+    const apiRequests = rig.readAllApiRequest();
+    const modelNames = apiRequests.map((r) => r.attributes?.model || 'unknown');
+
+    const proRequests = apiRequests.filter((r) =>
+      r.attributes?.model?.includes('pro'),
+    );
+    const flashRequests = apiRequests.filter((r) =>
+      r.attributes?.model?.includes('flash'),
+    );
+
+    expect(
+      proRequests.length,
+      `Expected at least one Pro request. Models used: ${modelNames.join(', ')}`,
+    ).toBeGreaterThanOrEqual(1);
+    expect(
+      flashRequests.length,
+      `Expected at least one Flash request after mode switch. Models used: ${modelNames.join(', ')}`,
+    ).toBeGreaterThanOrEqual(1);
+  });
 });
@@ -11,7 +11,7 @@
        "packages/*"
      ],
      "dependencies": {
-        "ink": "npm:@jrichman/ink@6.4.11",
+        "ink": "npm:@jrichman/ink@6.5.0",
        "latest-version": "^9.0.0",
        "node-fetch-native": "^1.6.7",
        "proper-lockfile": "^4.1.2",
@@ -10089,9 +10089,9 @@
    },
    "node_modules/ink": {
      "name": "@jrichman/ink",
-      "version": "6.4.11",
-      "resolved": "https://registry.npmjs.org/@jrichman/ink/-/ink-6.4.11.tgz",
-      "integrity": "sha512-93LQlzT7vvZ1XJcmOMwN4s+6W334QegendeHOMnEJBlhnpIzr8bws6/aOEHG8ZCuVD/vNeeea5m1msHIdAY6ig==",
+      "version": "6.5.0",
+      "resolved": "https://registry.npmjs.org/@jrichman/ink/-/ink-6.5.0.tgz",
+      "integrity": "sha512-S4g/ng7fPZmFwclO82iWkOce8vDLy/FIDgHIfkCWGOehqHe6dexHsmq3kNQD21okh198pA5SAQTCqNQJb/svRQ==",
      "license": "MIT",
      "dependencies": {
        "@alcalzone/ansi-tokenize": "^0.2.1",
@@ -10116,6 +10116,7 @@
        "type-fest": "^4.27.0",
        "wrap-ansi": "^9.0.0",
        "ws": "^8.18.0",
+        "yargs": "^17.7.2",
        "yoga-layout": "~3.2.1"
      },
      "engines": {
@@ -17550,7 +17551,7 @@
        "fzf": "^0.5.2",
        "glob": "^12.0.0",
        "highlight.js": "^11.11.1",
-        "ink": "npm:@jrichman/ink@6.4.11",
+        "ink": "npm:@jrichman/ink@6.5.0",
        "ink-gradient": "^3.0.0",
        "ink-spinner": "^5.0.0",
        "latest-version": "^9.0.0",
@@ -52,7 +52,7 @@
    "test:integration:sandbox:none": "cross-env GEMINI_SANDBOX=false vitest run --root ./integration-tests",
    "test:integration:sandbox:docker": "cross-env GEMINI_SANDBOX=docker npm run build:sandbox && cross-env GEMINI_SANDBOX=docker vitest run --root ./integration-tests",
    "test:integration:sandbox:podman": "cross-env GEMINI_SANDBOX=podman vitest run --root ./integration-tests",
-    "lint": "eslint . --cache --max-warnings 0",
+    "lint": "cross-env NODE_OPTIONS=\"--max-old-space-size=8192\" eslint . --cache --max-warnings 0",
    "lint:fix": "eslint . --fix --ext .ts,.tsx && eslint integration-tests --fix && eslint scripts --fix && npm run format",
    "lint:ci": "npm run lint:all",
    "lint:all": "node scripts/lint.js",
@@ -68,7 +68,7 @@
    "pre-commit": "node scripts/pre-commit.js"
  },
  "overrides": {
-    "ink": "npm:@jrichman/ink@6.4.11",
+    "ink": "npm:@jrichman/ink@6.5.0",
    "wrap-ansi": "9.0.2",
    "cliui": {
      "wrap-ansi": "7.0.0"
@@ -136,7 +136,7 @@
    "yargs": "^17.7.2"
  },
  "dependencies": {
-    "ink": "npm:@jrichman/ink@6.4.11",
+    "ink": "npm:@jrichman/ink@6.5.0",
    "latest-version": "^9.0.0",
    "node-fetch-native": "^1.6.7",
    "proper-lockfile": "^4.1.2",
@@ -352,23 +352,37 @@ describe('loadConfig', () => {
    });

    describe('interactivity', () => {
-      it('should set interactive true when not headless', async () => {
+      it('should always set interactive true', async () => {
+        vi.mocked(isHeadlessMode).mockReturnValue(true);
+        await loadConfig(mockSettings, mockExtensionLoader, taskId);
+        expect(Config).toHaveBeenCalledWith(
+          expect.objectContaining({
+            interactive: true,
+          }),
+        );
+
        vi.mocked(isHeadlessMode).mockReturnValue(false);
        await loadConfig(mockSettings, mockExtensionLoader, taskId);
        expect(Config).toHaveBeenCalledWith(
          expect.objectContaining({
            interactive: true,
-            enableInteractiveShell: true,
          }),
        );
      });

-      it('should set interactive false when headless', async () => {
+      it('should set enableInteractiveShell based on headless mode', async () => {
+        vi.mocked(isHeadlessMode).mockReturnValue(false);
+        await loadConfig(mockSettings, mockExtensionLoader, taskId);
+        expect(Config).toHaveBeenCalledWith(
+          expect.objectContaining({
+            enableInteractiveShell: true,
+          }),
+        );
+
        vi.mocked(isHeadlessMode).mockReturnValue(true);
        await loadConfig(mockSettings, mockExtensionLoader, taskId);
        expect(Config).toHaveBeenCalledWith(
          expect.objectContaining({
-            interactive: false,
            enableInteractiveShell: false,
          }),
        );
@@ -410,7 +424,22 @@ describe('loadConfig', () => {
      });
    });

-    describe('authentication fallback', () => {
+    describe('authentication logic', () => {
+      const setupConfigMock = (refreshAuthMock: ReturnType<typeof vi.fn>) => {
+        vi.mocked(Config).mockImplementation(
+          (params: unknown) =>
+            ({
+              ...(params as object),
+              initialize: vi.fn(),
+              waitForMcpInit: vi.fn(),
+              refreshAuth: refreshAuthMock,
+              getExperiments: vi.fn().mockReturnValue({ flags: {} }),
+              getRemoteAdminSettings: vi.fn(),
+              setRemoteAdminSettings: vi.fn(),
+            }) as unknown as Config,
+        );
+      };
+
      beforeEach(() => {
        vi.stubEnv('USE_CCPA', 'true');
        vi.stubEnv('GEMINI_API_KEY', '');
@@ -420,182 +449,77 @@ describe('loadConfig', () => {
        vi.unstubAllEnvs();
      });

-      it('should fall back to COMPUTE_ADC in Cloud Shell if LOGIN_WITH_GOOGLE fails', async () => {
-        vi.stubEnv('CLOUD_SHELL', 'true');
-        vi.mocked(isHeadlessMode).mockReturnValue(false);
-        const refreshAuthMock = vi.fn().mockImplementation((authType) => {
-          if (authType === AuthType.LOGIN_WITH_GOOGLE) {
-            throw new FatalAuthenticationError('Non-interactive session');
-          }
-          return Promise.resolve();
-        });
-
-        // Update the mock implementation for this test
-        vi.mocked(Config).mockImplementation(
-          (params: unknown) =>
-            ({
-              ...(params as object),
-              initialize: vi.fn(),
-              waitForMcpInit: vi.fn(),
-              refreshAuth: refreshAuthMock,
-              getExperiments: vi.fn().mockReturnValue({ flags: {} }),
-              getRemoteAdminSettings: vi.fn(),
-              setRemoteAdminSettings: vi.fn(),
-            }) as unknown as Config,
-        );
-
-        await loadConfig(mockSettings, mockExtensionLoader, taskId);
-
-        expect(refreshAuthMock).toHaveBeenCalledWith(
-          AuthType.LOGIN_WITH_GOOGLE,
-        );
-        expect(refreshAuthMock).toHaveBeenCalledWith(AuthType.COMPUTE_ADC);
-      });
-
-      it('should not fall back to COMPUTE_ADC if not in cloud environment', async () => {
-        vi.mocked(isHeadlessMode).mockReturnValue(false);
-        const refreshAuthMock = vi.fn().mockImplementation((authType) => {
-          if (authType === AuthType.LOGIN_WITH_GOOGLE) {
-            throw new FatalAuthenticationError('Non-interactive session');
-          }
-          return Promise.resolve();
-        });
-
-        vi.mocked(Config).mockImplementation(
-          (params: unknown) =>
-            ({
-              ...(params as object),
-              initialize: vi.fn(),
-              waitForMcpInit: vi.fn(),
-              refreshAuth: refreshAuthMock,
-              getExperiments: vi.fn().mockReturnValue({ flags: {} }),
-              getRemoteAdminSettings: vi.fn(),
-              setRemoteAdminSettings: vi.fn(),
-            }) as unknown as Config,
-        );
-
-        await expect(
-          loadConfig(mockSettings, mockExtensionLoader, taskId),
-        ).rejects.toThrow('Non-interactive session');
-
-        expect(refreshAuthMock).toHaveBeenCalledWith(
-          AuthType.LOGIN_WITH_GOOGLE,
-        );
-        expect(refreshAuthMock).not.toHaveBeenCalledWith(AuthType.COMPUTE_ADC);
-      });
-
-      it('should skip LOGIN_WITH_GOOGLE and use COMPUTE_ADC directly in headless Cloud Shell', async () => {
-        vi.stubEnv('CLOUD_SHELL', 'true');
-        vi.mocked(isHeadlessMode).mockReturnValue(true);
-
+      it('should attempt COMPUTE_ADC by default and bypass LOGIN_WITH_GOOGLE if successful', async () => {
        const refreshAuthMock = vi.fn().mockResolvedValue(undefined);
-
-        vi.mocked(Config).mockImplementation(
-          (params: unknown) =>
-            ({
-              ...(params as object),
-              initialize: vi.fn(),
-              waitForMcpInit: vi.fn(),
-              refreshAuth: refreshAuthMock,
-              getExperiments: vi.fn().mockReturnValue({ flags: {} }),
-              getRemoteAdminSettings: vi.fn(),
-              setRemoteAdminSettings: vi.fn(),
-            }) as unknown as Config,
-        );
+        setupConfigMock(refreshAuthMock);

        await loadConfig(mockSettings, mockExtensionLoader, taskId);

+        expect(refreshAuthMock).toHaveBeenCalledWith(AuthType.COMPUTE_ADC);
        expect(refreshAuthMock).not.toHaveBeenCalledWith(
          AuthType.LOGIN_WITH_GOOGLE,
        );
-        expect(refreshAuthMock).toHaveBeenCalledWith(AuthType.COMPUTE_ADC);
      });

-      it('should skip LOGIN_WITH_GOOGLE and use COMPUTE_ADC directly if GEMINI_CLI_USE_COMPUTE_ADC is true', async () => {
-        vi.stubEnv('GEMINI_CLI_USE_COMPUTE_ADC', 'true');
-        vi.mocked(isHeadlessMode).mockReturnValue(false); // Even if not headless
-
-        const refreshAuthMock = vi.fn().mockResolvedValue(undefined);
-
-        vi.mocked(Config).mockImplementation(
-          (params: unknown) =>
-            ({
-              ...(params as object),
-              initialize: vi.fn(),
-              waitForMcpInit: vi.fn(),
-              refreshAuth: refreshAuthMock,
-              getExperiments: vi.fn().mockReturnValue({ flags: {} }),
-              getRemoteAdminSettings: vi.fn(),
-              setRemoteAdminSettings: vi.fn(),
-            }) as unknown as Config,
-        );
+      it('should fallback to LOGIN_WITH_GOOGLE if COMPUTE_ADC fails and interactive mode is available', async () => {
+        vi.mocked(isHeadlessMode).mockReturnValue(false);
+        const refreshAuthMock = vi.fn().mockImplementation((authType) => {
+          if (authType === AuthType.COMPUTE_ADC) {
+            return Promise.reject(new Error('ADC failed'));
+          }
+          return Promise.resolve();
+        });
+        setupConfigMock(refreshAuthMock);

        await loadConfig(mockSettings, mockExtensionLoader, taskId);

-        expect(refreshAuthMock).not.toHaveBeenCalledWith(
+        expect(refreshAuthMock).toHaveBeenCalledWith(AuthType.COMPUTE_ADC);
+        expect(refreshAuthMock).toHaveBeenCalledWith(
          AuthType.LOGIN_WITH_GOOGLE,
        );
-        expect(refreshAuthMock).toHaveBeenCalledWith(AuthType.COMPUTE_ADC);
      });

-      it('should throw FatalAuthenticationError in headless mode if no ADC fallback available', async () => {
+      it('should throw FatalAuthenticationError in headless mode if COMPUTE_ADC fails', async () => {
        vi.mocked(isHeadlessMode).mockReturnValue(true);

-        const refreshAuthMock = vi.fn().mockResolvedValue(undefined);
-
-        vi.mocked(Config).mockImplementation(
-          (params: unknown) =>
-            ({
-              ...(params as object),
-              initialize: vi.fn(),
-              waitForMcpInit: vi.fn(),
-              refreshAuth: refreshAuthMock,
-              getExperiments: vi.fn().mockReturnValue({ flags: {} }),
-              getRemoteAdminSettings: vi.fn(),
-              setRemoteAdminSettings: vi.fn(),
-            }) as unknown as Config,
-        );
+        const refreshAuthMock = vi.fn().mockImplementation((authType) => {
+          if (authType === AuthType.COMPUTE_ADC) {
+            return Promise.reject(new Error('ADC not found'));
+          }
+          return Promise.resolve();
+        });
+        setupConfigMock(refreshAuthMock);

        await expect(
          loadConfig(mockSettings, mockExtensionLoader, taskId),
        ).rejects.toThrow(
-          'Interactive terminal required for LOGIN_WITH_GOOGLE. Run in an interactive terminal or set GEMINI_CLI_USE_COMPUTE_ADC=true to use Application Default Credentials.',
+          'COMPUTE_ADC failed: ADC not found. (LOGIN_WITH_GOOGLE fallback skipped due to headless mode. Run in an interactive terminal to use OAuth.)',
        );

-        expect(refreshAuthMock).not.toHaveBeenCalled();
+        expect(refreshAuthMock).toHaveBeenCalledWith(AuthType.COMPUTE_ADC);
+        expect(refreshAuthMock).not.toHaveBeenCalledWith(
+          AuthType.LOGIN_WITH_GOOGLE,
+        );
      });

-      it('should include both original and fallback error when COMPUTE_ADC fallback fails', async () => {
-        vi.stubEnv('CLOUD_SHELL', 'true');
+      it('should include both original and fallback error when LOGIN_WITH_GOOGLE fallback fails', async () => {
        vi.mocked(isHeadlessMode).mockReturnValue(false);

        const refreshAuthMock = vi.fn().mockImplementation((authType) => {
-          if (authType === AuthType.LOGIN_WITH_GOOGLE) {
-            throw new FatalAuthenticationError('OAuth failed');
-          }
          if (authType === AuthType.COMPUTE_ADC) {
            throw new Error('ADC failed');
          }
+          if (authType === AuthType.LOGIN_WITH_GOOGLE) {
+            throw new FatalAuthenticationError('OAuth failed');
+          }
          return Promise.resolve();
        });
-
-        vi.mocked(Config).mockImplementation(
-          (params: unknown) =>
-            ({
-              ...(params as object),
-              initialize: vi.fn(),
-              waitForMcpInit: vi.fn(),
-              refreshAuth: refreshAuthMock,
-              getExperiments: vi.fn().mockReturnValue({ flags: {} }),
-              getRemoteAdminSettings: vi.fn(),
-              setRemoteAdminSettings: vi.fn(),
-            }) as unknown as Config,
-        );
+        setupConfigMock(refreshAuthMock);

        await expect(
          loadConfig(mockSettings, mockExtensionLoader, taskId),
        ).rejects.toThrow(
-          'OAuth failed. Fallback to COMPUTE_ADC also failed: ADC failed',
+          'OAuth failed. The initial COMPUTE_ADC attempt also failed: ADC failed',
        );
      });
    });
@@ -25,7 +25,6 @@ import {
  ExperimentFlags,
  isHeadlessMode,
  FatalAuthenticationError,
-  isCloudShell,
  PolicyDecision,
  PRIORITY_YOLO_ALLOW_ALL,
  type TelemetryTarget,
@@ -43,7 +42,6 @@ export async function loadConfig(
  taskId: string,
 ): Promise<Config> {
  const workspaceDir = process.cwd();
-  const adcFilePath = process.env['GOOGLE_APPLICATION_CREDENTIALS'];

  const folderTrust =
    settings.folderTrust === true ||
@@ -125,7 +123,7 @@ export async function loadConfig(
    trustedFolder: true,
    extensionLoader,
    checkpointing,
-    interactive: !isHeadlessMode(),
+    interactive: true,
    enableInteractiveShell: !isHeadlessMode(),
    ptyInfo: 'auto',
    enableAgents: settings.experimental?.enableAgents ?? true,
@@ -192,7 +190,7 @@ export async function loadConfig(
  await config.waitForMcpInit();
  startupProfiler.flush(config);

-  await refreshAuthentication(config, adcFilePath, 'Config');
+  await refreshAuthentication(config, 'Config');

  return config;
 }
@@ -263,75 +261,51 @@ function findEnvFile(startDir: string): string | null {

 async function refreshAuthentication(
  config: Config,
-  adcFilePath: string | undefined,
  logPrefix: string,
 ): Promise<void> {
  if (process.env['USE_CCPA']) {
    logger.info(`[${logPrefix}] Using CCPA Auth:`);
+
+    logger.info(`[${logPrefix}] Attempting COMPUTE_ADC first.`);
    try {
-      if (adcFilePath) {
-        path.resolve(adcFilePath);
-      }
-    } catch (e) {
-      logger.error(
-        `[${logPrefix}] USE_CCPA env var is true but unable to resolve GOOGLE_APPLICATION_CREDENTIALS file path ${adcFilePath}. Error ${e}`,
+      await config.refreshAuth(AuthType.COMPUTE_ADC);
+      logger.info(`[${logPrefix}] COMPUTE_ADC successful.`);
+    } catch (adcError) {
+      const adcMessage =
+        adcError instanceof Error ? adcError.message : String(adcError);
+      logger.info(
+        `[${logPrefix}] COMPUTE_ADC failed or not available: ${adcMessage}`,
      );
-    }

-    const useComputeAdc = process.env['GEMINI_CLI_USE_COMPUTE_ADC'] === 'true';
-    const isHeadless = isHeadlessMode();
-    const shouldSkipOauth = isHeadless || useComputeAdc;
+      const useComputeAdc =
+        process.env['GEMINI_CLI_USE_COMPUTE_ADC'] === 'true';
+      const isHeadless = isHeadlessMode();

-    if (shouldSkipOauth) {
-      if (isCloudShell() || useComputeAdc) {
-        logger.info(
-          `[${logPrefix}] Skipping LOGIN_WITH_GOOGLE due to ${isHeadless ? 'headless mode' : 'GEMINI_CLI_USE_COMPUTE_ADC'}. Attempting COMPUTE_ADC.`,
-        );
-        try {
-          await config.refreshAuth(AuthType.COMPUTE_ADC);
-          logger.info(`[${logPrefix}] COMPUTE_ADC successful.`);
-        } catch (adcError) {
-          const adcMessage =
-            adcError instanceof Error ? adcError.message : String(adcError);
-          throw new FatalAuthenticationError(
-            `COMPUTE_ADC failed: ${adcMessage}. (Skipped LOGIN_WITH_GOOGLE due to ${isHeadless ? 'headless mode' : 'GEMINI_CLI_USE_COMPUTE_ADC'})`,
-          );
-        }
-      } else {
+      if (isHeadless || useComputeAdc) {
+        const reason = isHeadless
+          ? 'headless mode'
+          : 'GEMINI_CLI_USE_COMPUTE_ADC=true';
        throw new FatalAuthenticationError(
-          `Interactive terminal required for LOGIN_WITH_GOOGLE. Run in an interactive terminal or set GEMINI_CLI_USE_COMPUTE_ADC=true to use Application Default Credentials.`,
+          `COMPUTE_ADC failed: ${adcMessage}. (LOGIN_WITH_GOOGLE fallback skipped due to ${reason}. Run in an interactive terminal to use OAuth.)`,
        );
      }
-    } else {
+
+      logger.info(
+        `[${logPrefix}] COMPUTE_ADC failed, falling back to LOGIN_WITH_GOOGLE.`,
+      );
      try {
        await config.refreshAuth(AuthType.LOGIN_WITH_GOOGLE);
      } catch (e) {
-        if (
-          e instanceof FatalAuthenticationError &&
-          (isCloudShell() || useComputeAdc)
-        ) {
-          logger.warn(
-            `[${logPrefix}] LOGIN_WITH_GOOGLE failed. Attempting COMPUTE_ADC fallback.`,
+        if (e instanceof FatalAuthenticationError) {
+          const originalMessage = e instanceof Error ? e.message : String(e);
+          throw new FatalAuthenticationError(
+            `${originalMessage}. The initial COMPUTE_ADC attempt also failed: ${adcMessage}`,
          );
-          try {
-            await config.refreshAuth(AuthType.COMPUTE_ADC);
-            logger.info(`[${logPrefix}] COMPUTE_ADC fallback successful.`);
-          } catch (adcError) {
-            logger.error(
-              `[${logPrefix}] COMPUTE_ADC fallback failed: ${adcError}`,
-            );
-            const originalMessage = e instanceof Error ? e.message : String(e);
-            const adcMessage =
-              adcError instanceof Error ? adcError.message : String(adcError);
-            throw new FatalAuthenticationError(
-              `${originalMessage}. Fallback to COMPUTE_ADC also failed: ${adcMessage}`,
-            );
-          }
-        } else {
-          throw e;
        }
+        throw e;
      }
    }
+
    logger.info(
      `[${logPrefix}] GOOGLE_CLOUD_PROJECT: ${process.env['GOOGLE_CLOUD_PROJECT']}`,
    );
@@ -109,6 +109,12 @@ export function createMockConfig(
        enableEnvironmentVariableRedaction: false,
      },
    }),
+    isExperimentalAgentHistoryTruncationEnabled: vi.fn().mockReturnValue(false),
+    getExperimentalAgentHistoryTruncationThreshold: vi.fn().mockReturnValue(50),
+    getExperimentalAgentHistoryRetainedMessages: vi.fn().mockReturnValue(30),
+    isExperimentalAgentHistorySummarizationEnabled: vi
+      .fn()
+      .mockReturnValue(false),
    ...overrides,
  } as unknown as Config;

@@ -49,7 +49,7 @@
    "fzf": "^0.5.2",
    "glob": "^12.0.0",
    "highlight.js": "^11.11.1",
-    "ink": "npm:@jrichman/ink@6.4.11",
+    "ink": "npm:@jrichman/ink@6.5.0",
    "ink-gradient": "^3.0.0",
    "ink-spinner": "^5.0.0",
    "latest-version": "^9.0.0",
@@ -28,6 +28,7 @@ import {
  LlmRole,
  type GitService,
  processSingleFileContent,
+  InvalidStreamError,
 } from '@google/gemini-cli-core';
 import {
  SettingScope,
@@ -99,6 +100,8 @@ vi.mock(
    const actual = await importOriginal();
    return {
      ...actual,
+      updatePolicy: vi.fn(),
+      createPolicyUpdater: vi.fn(),
      ReadManyFilesTool: vi.fn().mockImplementation(() => ({
        name: 'read_many_files',
        kind: 'read',
@@ -181,6 +184,20 @@ describe('GeminiAgent', () => {
      getWorkspaceContext: vi.fn().mockReturnValue({
        addReadOnlyPath: vi.fn(),
      }),
+      getPolicyEngine: vi.fn().mockReturnValue({
+        addRule: vi.fn(),
+      }),
+      messageBus: {
+        publish: vi.fn(),
+        subscribe: vi.fn(),
+        unsubscribe: vi.fn(),
+      },
+      storage: {
+        getWorkspaceAutoSavedPolicyPath: vi.fn(),
+        getAutoSavedPolicyPath: vi.fn(),
+        setClientName: vi.fn(),
+      },
+      setClientName: vi.fn(),
      get config() {
        return this;
      },
@@ -201,7 +218,10 @@ describe('GeminiAgent', () => {
    (loadCliConfig as unknown as Mock).mockResolvedValue(mockConfig);
    (loadSettings as unknown as Mock).mockImplementation(() => ({
      merged: {
-        security: { auth: { selectedType: AuthType.LOGIN_WITH_GOOGLE } },
+        security: {
+          auth: { selectedType: AuthType.LOGIN_WITH_GOOGLE },
+          enablePermanentToolApproval: true,
+        },
        mcpServers: {},
      },
      setValue: vi.fn(),
@@ -687,7 +707,10 @@ describe('Session', () => {
      systemDefaults: { settings: {} },
      user: { settings: {} },
      workspace: { settings: {} },
-      merged: { settings: {} },
+      merged: {
+        security: { enablePermanentToolApproval: true },
+        mcpServers: {},
+      },
      errors: [],
    } as unknown as LoadedSettings);
  });
@@ -763,6 +786,32 @@ describe('Session', () => {
    expect(result).toMatchObject({ stopReason: 'end_turn' });
  });

+  it('should handle prompt with empty response (InvalidStreamError)', async () => {
+    mockChat.sendMessageStream.mockRejectedValue(
+      new InvalidStreamError('Empty response', 'NO_RESPONSE_TEXT'),
+    );
+
+    const result = await session.prompt({
+      sessionId: 'session-1',
+      prompt: [{ type: 'text', text: 'Hi' }],
+    });
+
+    expect(mockChat.sendMessageStream).toHaveBeenCalled();
+    expect(result).toMatchObject({ stopReason: 'end_turn' });
+  });
+
+  it('should handle prompt with empty response (NO_RESPONSE_TEXT anomaly)', async () => {
+    mockChat.sendMessageStream.mockRejectedValue({ type: 'NO_RESPONSE_TEXT' });
+
+    const result = await session.prompt({
+      sessionId: 'session-1',
+      prompt: [{ type: 'text', text: 'Hi' }],
+    });
+
+    expect(mockChat.sendMessageStream).toHaveBeenCalled();
+    expect(result).toMatchObject({ stopReason: 'end_turn' });
+  });
+
  it('should handle /memory command', async () => {
    const handleCommandSpy = vi
      .spyOn(
@@ -1026,6 +1075,166 @@ describe('Session', () => {
    );
  });

+  it('should exclude always allow and save permanent option when enablePermanentToolApproval is false', async () => {
+    mockConfig.getDisableAlwaysAllow = vi.fn().mockReturnValue(false);
+    const confirmationDetails = {
+      type: 'edit',
+      onConfirm: vi.fn(),
+    };
+    mockTool.build.mockReturnValue({
+      getDescription: () => 'Test Tool',
+      toolLocations: () => [],
+      shouldConfirmExecute: vi.fn().mockResolvedValue(confirmationDetails),
+      execute: vi.fn().mockResolvedValue({ llmContent: 'Tool Result' }),
+    });
+
+    const customSettings = {
+      system: { settings: {} },
+      systemDefaults: { settings: {} },
+      user: { settings: {} },
+      workspace: { settings: {} },
+      merged: {
+        security: { enablePermanentToolApproval: false },
+        mcpServers: {},
+      },
+      errors: [],
+    } as unknown as LoadedSettings;
+
+    const localSession = new Session(
+      'session-2',
+      mockChat,
+      mockConfig,
+      mockConnection,
+      customSettings,
+    );
+
+    mockConnection.requestPermission.mockResolvedValueOnce({
+      outcome: {
+        outcome: 'selected',
+        optionId: ToolConfirmationOutcome.ProceedOnce,
+      },
+    });
+
+    const stream1 = createMockStream([
+      {
+        type: StreamEventType.CHUNK,
+        value: {
+          functionCalls: [{ name: 'test_tool', args: {} }],
+        },
+      },
+    ]);
+    const stream2 = createMockStream([
+      {
+        type: StreamEventType.CHUNK,
+        value: { candidates: [] },
+      },
+    ]);
+
+    mockChat.sendMessageStream
+      .mockResolvedValueOnce(stream1)
+      .mockResolvedValueOnce(stream2);
+
+    await localSession.prompt({
+      sessionId: 'session-2',
+      prompt: [{ type: 'text', text: 'Call tool' }],
+    });
+
+    expect(mockConnection.requestPermission).toHaveBeenCalledWith(
+      expect.objectContaining({
+        options: expect.not.arrayContaining([
+          expect.objectContaining({
+            optionId: ToolConfirmationOutcome.ProceedAlwaysAndSave,
+          }),
+        ]),
+      }),
+    );
+    expect(mockConnection.requestPermission).toHaveBeenCalledWith(
+      expect.objectContaining({
+        options: expect.arrayContaining([
+          expect.objectContaining({
+            optionId: ToolConfirmationOutcome.ProceedAlways,
+          }),
+        ]),
+      }),
+    );
+  });
+
+  it('should include always allow and save permanent option when enablePermanentToolApproval is true', async () => {
+    mockConfig.getDisableAlwaysAllow = vi.fn().mockReturnValue(false);
+    const confirmationDetails = {
+      type: 'edit',
+      onConfirm: vi.fn(),
+    };
+    mockTool.build.mockReturnValue({
+      getDescription: () => 'Test Tool',
+      toolLocations: () => [],
+      shouldConfirmExecute: vi.fn().mockResolvedValue(confirmationDetails),
+      execute: vi.fn().mockResolvedValue({ llmContent: 'Tool Result' }),
+    });
+
+    const customSettings = {
+      system: { settings: {} },
+      systemDefaults: { settings: {} },
+      user: { settings: {} },
+      workspace: { settings: {} },
+      merged: {
+        security: { enablePermanentToolApproval: true },
+        mcpServers: {},
+      },
+      errors: [],
+    } as unknown as LoadedSettings;
+
+    const localSession = new Session(
+      'session-2',
+      mockChat,
+      mockConfig,
+      mockConnection,
+      customSettings,
+    );
+
+    mockConnection.requestPermission.mockResolvedValueOnce({
+      outcome: {
+        outcome: 'selected',
+        optionId: ToolConfirmationOutcome.ProceedOnce,
+      },
+    });
+
+    const stream1 = createMockStream([
+      {
+        type: StreamEventType.CHUNK,
+        value: {
+          functionCalls: [{ name: 'test_tool', args: {} }],
+        },
+      },
+    ]);
+    const stream2 = createMockStream([
+      {
+        type: StreamEventType.CHUNK,
+        value: { candidates: [] },
+      },
+    ]);
+
+    mockChat.sendMessageStream
+      .mockResolvedValueOnce(stream1)
+      .mockResolvedValueOnce(stream2);
+
+    await localSession.prompt({
+      sessionId: 'session-2',
+      prompt: [{ type: 'text', text: 'Call tool' }],
+    });
+
+    expect(mockConnection.requestPermission).toHaveBeenCalledWith(
+      expect.objectContaining({
+        options: expect.arrayContaining([
+          expect.objectContaining({
+            optionId: ToolConfirmationOutcome.ProceedAlwaysAndSave,
+            name: 'Allow for this file in all future sessions',
+          }),
+        ]),
+      }),
+    );
+  });
+
  it('should use filePath for ACP diff content in permission request', async () => {
    const confirmationDetails = {
      type: 'edit',
@@ -1154,6 +1363,56 @@ describe('Session', () => {
    );
  });

+  it('should call updatePolicy when tool permission triggers always allow', async () => {
+    const confirmationDetails = {
+      type: 'info',
+      onConfirm: vi.fn(),
+    };
+    mockTool.build.mockReturnValue({
+      getDescription: () => 'Test Tool',
+      toolLocations: () => [],
+      shouldConfirmExecute: vi.fn().mockResolvedValue(confirmationDetails),
+      execute: vi.fn().mockResolvedValue({ llmContent: 'Tool Result' }),
+    });
+
+    mockConnection.requestPermission.mockResolvedValue({
+      outcome: {
+        outcome: 'selected',
+        optionId: ToolConfirmationOutcome.ProceedAlways,
+      },
+    });
+
+    const stream1 = createMockStream([
+      {
+        type: StreamEventType.CHUNK,
+        value: {
+          functionCalls: [{ name: 'test_tool', args: {} }],
+        },
+      },
+    ]);
+    const stream2 = createMockStream([
+      {
+        type: StreamEventType.CHUNK,
+        value: { candidates: [] },
+      },
+    ]);
+
+    mockChat.sendMessageStream
+      .mockResolvedValueOnce(stream1)
+      .mockResolvedValueOnce(stream2);
+
+    const { updatePolicy } = await import('@google/gemini-cli-core');
+
+    await session.prompt({
+      sessionId: 'session-1',
+      prompt: [{ type: 'text', text: 'Call tool' }],
+    });
+
+    expect(confirmationDetails.onConfirm).toHaveBeenCalled();
+
+    expect(updatePolicy).toHaveBeenCalled();
+  });
+
  it('should use filePath for ACP diff content in tool result', async () => {
    mockTool.build.mockReturnValue({
      getDescription: () => 'Test Tool',
@@ -48,7 +48,9 @@ import {
  PREVIEW_GEMINI_MODEL_AUTO,
  getDisplayString,
  processSingleFileContent,
+  InvalidStreamError,
  type AgentLoopContext,
+  updatePolicy,
 } from '@google/gemini-cli-core';
 import * as acp from '@agentclientprotocol/sdk';
 import { AcpFileSystemService } from './fileSystemService.js';
@@ -64,6 +66,7 @@ import {
  loadSettings,
  type LoadedSettings,
 } from '../config/settings.js';
+import { createPolicyUpdater } from '../config/policy.js';
 import * as fs from 'node:fs/promises';
 import * as path from 'node:path';
 import { z } from 'zod';
@@ -133,6 +136,7 @@ export class GeminiAgent {
    args: acp.InitializeRequest,
  ): Promise<acp.InitializeResponse> {
    this.clientCapabilities = args.clientCapabilities;
+
    const authMethods = [
      {
        id: AuthType.LOGIN_WITH_GOOGLE,
@@ -322,6 +326,7 @@ export class GeminiAgent {

    const geminiClient = config.getGeminiClient();
    const chat = await geminiClient.startChat();
+
    const session = new Session(
      sessionId,
      chat,
@@ -512,6 +517,12 @@ export class GeminiAgent {

    const config = await loadCliConfig(settings, sessionId, this.argv, { cwd });

+    createPolicyUpdater(
+      config.getPolicyEngine(),
+      config.messageBus,
+      config.storage,
+    );
+
    return config;
  }

@@ -841,6 +852,37 @@ export class Session {
          return { stopReason: CoreToolCallStatus.Cancelled };
        }

+        if (
+          error instanceof InvalidStreamError ||
+          (error &&
+            typeof error === 'object' &&
+            'type' in error &&
+            error.type === 'NO_RESPONSE_TEXT')
+        ) {
+          // The stream ended with an empty response or malformed tool call.
+          // Treat this as a graceful end to the model's turn rather than a crash.
+          return {
+            stopReason: 'end_turn',
+            _meta: {
+              quota: {
+                token_count: {
+                  input_tokens: totalInputTokens,
+                  output_tokens: totalOutputTokens,
+                },
+                model_usage: Array.from(modelUsageMap.entries()).map(
+                  ([modelName, counts]) => ({
+                    model: modelName,
+                    token_count: {
+                      input_tokens: counts.input,
+                      output_tokens: counts.output,
+                    },
+                  }),
+                ),
+              },
+            },
+          };
+        }
+
        throw new acp.RequestError(
          getErrorStatus(error) || 500,
          getAcpErrorMessage(error),
@@ -1012,6 +1054,7 @@ export class Session {
          options: toPermissionOptions(
            confirmationDetails,
            this.context.config,
+            this.settings.merged.security.enablePermanentToolApproval,
          ),
          toolCall: {
            toolCallId: callId,
@@ -1036,6 +1079,16 @@ export class Session {

        await confirmationDetails.onConfirm(outcome);

+        // Update policy to enable Always Allow persistence
+        await updatePolicy(
+          tool,
+          outcome,
+          confirmationDetails,
+          this.context,
+          this.context.messageBus,
+          invocation,
+        );
+
        switch (outcome) {
          case ToolConfirmationOutcome.Cancel:
            return errorResponse(
@@ -1785,6 +1838,7 @@ const basicPermissionOptions = [
 function toPermissionOptions(
  confirmation: ToolCallConfirmationDetails,
  config: Config,
+  enablePermanentToolApproval: boolean = false,
 ): acp.PermissionOption[] {
  const disableAlwaysAllow = config.getDisableAlwaysAllow();
  const options: acp.PermissionOption[] = [];
@@ -1794,37 +1848,65 @@ function toPermissionOptions(
      case 'edit':
        options.push({
          optionId: ToolConfirmationOutcome.ProceedAlways,
-          name: 'Allow All Edits',
+          name: 'Allow for this session',
          kind: 'allow_always',
        });
+        if (enablePermanentToolApproval) {
+          options.push({
+            optionId: ToolConfirmationOutcome.ProceedAlwaysAndSave,
+            name: 'Allow for this file in all future sessions',
+            kind: 'allow_always',
+          });
+        }
        break;
      case 'exec':
        options.push({
          optionId: ToolConfirmationOutcome.ProceedAlways,
-          name: `Always Allow ${confirmation.rootCommand}`,
+          name: 'Allow for this session',
          kind: 'allow_always',
        });
+        if (enablePermanentToolApproval) {
+          options.push({
+            optionId: ToolConfirmationOutcome.ProceedAlwaysAndSave,
+            name: 'Allow this command for all future sessions',
+            kind: 'allow_always',
+          });
+        }
        break;
      case 'mcp':
        options.push(
          {
            optionId: ToolConfirmationOutcome.ProceedAlwaysServer,
-            name: `Always Allow ${confirmation.serverName}`,
+            name: 'Allow all server tools for this session',
            kind: 'allow_always',
          },
          {
            optionId: ToolConfirmationOutcome.ProceedAlwaysTool,
-            name: `Always Allow ${confirmation.toolName}`,
+            name: 'Allow tool for this session',
            kind: 'allow_always',
          },
        );
+        if (enablePermanentToolApproval) {
+          options.push({
+            optionId: ToolConfirmationOutcome.ProceedAlwaysAndSave,
+            name: 'Allow tool for all future sessions',
+            kind: 'allow_always',
+          });
+        }
        break;
      case 'info':
        options.push({
          optionId: ToolConfirmationOutcome.ProceedAlways,
-          name: `Always Allow`,
+          name: 'Allow for this session',
          kind: 'allow_always',
        });
+        if (enablePermanentToolApproval) {
+          options.push({
+            optionId: ToolConfirmationOutcome.ProceedAlwaysAndSave,
+            name: 'Allow for all future sessions',
+            kind: 'allow_always',
+          });
+        }
        break;
      case 'ask_user':
      case 'exit_plan_mode':
@@ -91,6 +91,14 @@ describe('GeminiAgent Session Resume', () => {
      storage: {
        getProjectTempDir: vi.fn().mockReturnValue('/tmp/project'),
      },
+      getPolicyEngine: vi.fn().mockReturnValue({
+        addRule: vi.fn(),
+      }),
+      messageBus: {
+        publish: vi.fn(),
+        subscribe: vi.fn(),
+        unsubscribe: vi.fn(),
+      },
      getApprovalMode: vi.fn().mockReturnValue('default'),
      isPlanEnabled: vi.fn().mockReturnValue(true),
      getModel: vi.fn().mockReturnValue('gemini-pro'),
@@ -143,12 +143,17 @@ vi.mock('@google/gemini-cli-core', async () => {
      respectGeminiIgnore: true,
      customIgnoreFilePaths: [],
    },
-    createPolicyEngineConfig: vi.fn(async () => ({
-      rules: [],
-      checkers: [],
-      defaultDecision: ServerConfig.PolicyDecision.ASK_USER,
-      approvalMode: ServerConfig.ApprovalMode.DEFAULT,
-    })),
+    createPolicyEngineConfig: vi.fn(
+      async (_settings, approvalMode, _workspacePoliciesDir, interactive) => ({
+        rules: [],
+        checkers: [],
+        defaultDecision: interactive
+          ? ServerConfig.PolicyDecision.ASK_USER
+          : ServerConfig.PolicyDecision.DENY,
+        approvalMode: approvalMode ?? ServerConfig.ApprovalMode.DEFAULT,
+        nonInteractive: !interactive,
+      }),
+    ),
    getAdminErrorMessage: vi.fn(
      (_feature) =>
        `YOLO mode is disabled by your administrator. To enable it, please request an update to the settings at: https://goo.gle/manage-gemini-cli`,
@@ -984,6 +989,7 @@ describe('Hierarchical Memory Loading (config.ts) - Placeholder Suite', () => {
        respectGeminiIgnore: true,
      }),
      200, // maxDirs
+      ['.git'], // boundaryMarkers
    );
  });

@@ -1013,6 +1019,7 @@ describe('Hierarchical Memory Loading (config.ts) - Placeholder Suite', () => {
        respectGeminiIgnore: true,
      }),
      200,
+      ['.git'], // boundaryMarkers
    );
  });

@@ -1041,6 +1048,7 @@ describe('Hierarchical Memory Loading (config.ts) - Placeholder Suite', () => {
        respectGeminiIgnore: true,
      }),
      200,
+      ['.git'], // boundaryMarkers
    );
  });
 });
@@ -1117,12 +1125,7 @@ describe('mergeExcludeTools', () => {
    ]);
    process.argv = ['node', 'script.js'];
    const argv = await parseArguments(createTestMergedSettings());
-    const config = await loadCliConfig(
-      settings,
-
-      'test-session',
-      argv,
-    );
+    const config = await loadCliConfig(settings, 'test-session', argv);
    expect(config.getExcludeTools()).toEqual(
      new Set(['tool1', 'tool2', 'tool3', 'tool4', 'tool5']),
    );
@@ -3460,6 +3463,8 @@ describe('Policy Engine Integration in loadCliConfig', () => {
        }),
      }),
      expect.anything(),
+      undefined,
+      expect.anything(),
    );
  });

@@ -3481,6 +3486,8 @@ describe('Policy Engine Integration in loadCliConfig', () => {
        }),
      }),
      expect.anything(),
+      undefined,
+      expect.anything(),
    );
  });

@@ -3504,6 +3511,8 @@ describe('Policy Engine Integration in loadCliConfig', () => {
        ],
      }),
      expect.anything(),
+      undefined,
+      expect.anything(),
    );
  });
 });
@@ -642,6 +642,7 @@ export async function loadCliConfig(
      memoryImportFormat,
      memoryFileFiltering,
      settings.context?.discoveryMaxDirs,
+      settings.context?.memoryBoundaryMarkers,
    );
    memoryContent = result.memoryContent;
    fileCount = result.fileCount;
@@ -792,8 +793,8 @@ export async function loadCliConfig(
    effectiveSettings,
    approvalMode,
    workspacePoliciesDir,
+    interactive,
  );
-  policyEngineConfig.nonInteractive = !interactive;

  const defaultModel = PREVIEW_GEMINI_MODEL_AUTO;
  const specifiedModel =
@@ -896,6 +897,7 @@ export async function loadCliConfig(
    loadMemoryFromIncludeDirectories:
      settings.context?.loadMemoryFromIncludeDirectories || false,
    discoveryMaxDirs: settings.context?.discoveryMaxDirs,
+    memoryBoundaryMarkers: settings.context?.memoryBoundaryMarkers,
    importFormat: settings.context?.importFormat,
    debugMode,
    question,
@@ -975,9 +977,18 @@ export async function loadCliConfig(
    disabledSkills: settings.skills?.disabled,
    experimentalJitContext: settings.experimental?.jitContext,
    experimentalMemoryManager: settings.experimental?.memoryManager,
+    experimentalAgentHistoryTruncation:
+      settings.experimental?.agentHistoryTruncation,
+    experimentalAgentHistoryTruncationThreshold:
+      settings.experimental?.agentHistoryTruncationThreshold,
+    experimentalAgentHistoryRetainedMessages:
+      settings.experimental?.agentHistoryRetainedMessages,
+    experimentalAgentHistorySummarization:
+      settings.experimental?.agentHistorySummarization,
    modelSteering: settings.experimental?.modelSteering,
    topicUpdateNarration: settings.experimental?.topicUpdateNarration,
    toolOutputMasking: settings.experimental?.toolOutputMasking,
+    useAgentProtocol: settings.experimental?.useAgentProtocol,
    noBrowser: !!process.env['NO_BROWSER'],
    summarizeToolOutput: settings.model?.summarizeToolOutput,
    ideMode,
@@ -990,6 +1001,8 @@ export async function loadCliConfig(
    useAlternateBuffer: settings.ui?.useAlternateBuffer,
    useRipgrep: settings.tools?.useRipgrep,
    enableInteractiveShell: settings.tools?.shell?.enableInteractiveShell,
+    shellBackgroundCompletionBehavior: settings.tools?.shell
+      ?.backgroundCompletionBehavior as string | undefined,
    shellToolInactivityTimeout: settings.tools?.shell?.inactivityTimeout,
    enableShellOutputEfficiency:
      settings.tools?.shell?.enableShellOutputEfficiency ?? true,
@@ -199,6 +199,7 @@ describe('ExtensionManager theme loading', () => {
        respectGeminiIgnore: true,
      }),
      getDiscoveryMaxDirs: () => 200,
+      getMemoryBoundaryMarkers: () => ['.git'],
      getMcpClientManager: () => ({
        getMcpInstructions: () => '',
        startExtension: vi.fn().mockResolvedValue(undefined),
@@ -605,12 +605,12 @@ describe('Policy Engine Integration Tests', () => {
    it('should verify non-interactive mode transformation', async () => {
      const settings: Settings = {};

-      const config = await createPolicyEngineConfig(
+      const engineConfig = await createPolicyEngineConfig(
        settings,
        ApprovalMode.DEFAULT,
+        undefined,
+        false,
      );
-      // Enable non-interactive mode
-      const engineConfig = { ...config, nonInteractive: true };
      const engine = new PolicyEngine(engineConfig);

      // ASK_USER should become DENY in non-interactive mode
@@ -53,6 +53,7 @@ export async function createPolicyEngineConfig(
  settings: Settings,
  approvalMode: ApprovalMode,
  workspacePoliciesDir?: string,
+  interactive: boolean = true,
 ): Promise<PolicyEngineConfig> {
  // Explicitly construct PolicySettings from Settings to ensure type safety
  // and avoid accidental leakage of other settings properties.
@@ -68,7 +69,12 @@ export async function createPolicyEngineConfig(
      settings.admin?.secureModeEnabled,
  };

-  return createCorePolicyEngineConfig(policySettings, approvalMode);
+  return createCorePolicyEngineConfig(
+    policySettings,
+    approvalMode,
+    undefined,
+    interactive,
+  );
 }

 export function createPolicyUpdater(
@@ -93,7 +93,7 @@ describe('loadSandboxConfig', () => {
      expect(config).toEqual({
        enabled: true,
        allowedPaths: [],
-        networkAccess: false,
+        networkAccess: true,
        command: 'docker',
        image: 'default/image',
      });
@@ -122,7 +122,7 @@ describe('loadSandboxConfig', () => {
      expect(config).toEqual({
        enabled: true,
        allowedPaths: [],
-        networkAccess: false,
+        networkAccess: true,
        command: 'lxc',
        image: 'default/image',
      });
@@ -148,7 +148,7 @@ describe('loadSandboxConfig', () => {
      expect(config).toEqual({
        enabled: true,
        allowedPaths: [],
-        networkAccess: false,
+        networkAccess: true,
        command: 'sandbox-exec',
        image: 'default/image',
      });
@@ -161,7 +161,7 @@ describe('loadSandboxConfig', () => {
      expect(config).toEqual({
        enabled: true,
        allowedPaths: [],
-        networkAccess: false,
+        networkAccess: true,
        command: 'sandbox-exec',
        image: 'default/image',
      });
@@ -174,7 +174,7 @@ describe('loadSandboxConfig', () => {
      expect(config).toEqual({
        enabled: true,
        allowedPaths: [],
-        networkAccess: false,
+        networkAccess: true,
        command: 'docker',
        image: 'default/image',
      });
@@ -187,7 +187,7 @@ describe('loadSandboxConfig', () => {
      expect(config).toEqual({
        enabled: true,
        allowedPaths: [],
-        networkAccess: false,
+        networkAccess: true,
        command: 'podman',
        image: 'default/image',
      });
@@ -210,7 +210,7 @@ describe('loadSandboxConfig', () => {
      expect(config).toEqual({
        enabled: true,
        allowedPaths: [],
-        networkAccess: false,
+        networkAccess: true,
        command: 'podman',
        image: 'default/image',
      });
@@ -244,7 +244,7 @@ describe('loadSandboxConfig', () => {
      expect(config).toEqual({
        enabled: true,
        allowedPaths: [],
-        networkAccess: false,
+        networkAccess: true,
        command: 'docker',
        image: 'env/image',
      });
@@ -257,7 +257,7 @@ describe('loadSandboxConfig', () => {
      expect(config).toEqual({
        enabled: true,
        allowedPaths: [],
-        networkAccess: false,
+        networkAccess: true,
        command: 'docker',
        image: 'default/image',
      });
@@ -285,7 +285,7 @@ describe('loadSandboxConfig', () => {
        expect(config).toEqual({
          enabled: true,
          allowedPaths: [],
-          networkAccess: false,
+          networkAccess: true,
          command: 'docker',
          image: 'default/image',
        });
@@ -339,7 +339,7 @@ describe('loadSandboxConfig', () => {
              enabled: true,
              command: 'podman',
              allowedPaths: [],
-              networkAccess: false,
+              networkAccess: true,
            },
          },
        },
@@ -356,7 +356,7 @@ describe('loadSandboxConfig', () => {
              enabled: true,
              image: 'custom/image',
              allowedPaths: [],
-              networkAccess: false,
+              networkAccess: true,
            },
          },
        },
@@ -372,7 +372,7 @@ describe('loadSandboxConfig', () => {
            sandbox: {
              enabled: false,
              allowedPaths: [],
-              networkAccess: false,
+              networkAccess: true,
            },
          },
        },
@@ -388,7 +388,7 @@ describe('loadSandboxConfig', () => {
            sandbox: {
              enabled: true,
              allowedPaths: ['/settings-path'],
-              networkAccess: false,
+              networkAccess: true,
            },
          },
        },
@@ -410,7 +410,7 @@ describe('loadSandboxConfig', () => {
      expect(config).toEqual({
        enabled: true,
        allowedPaths: [],
-        networkAccess: false,
+        networkAccess: true,
        command: 'runsc',
        image: 'default/image',
      });
@@ -425,7 +425,7 @@ describe('loadSandboxConfig', () => {
      expect(config).toEqual({
        enabled: true,
        allowedPaths: [],
-        networkAccess: false,
+        networkAccess: true,
        command: 'runsc',
        image: 'default/image',
      });
@@ -442,7 +442,7 @@ describe('loadSandboxConfig', () => {
      expect(config).toEqual({
        enabled: true,
        allowedPaths: [],
-        networkAccess: false,
+        networkAccess: true,
        command: 'runsc',
        image: 'default/image',
      });
@@ -460,7 +460,7 @@ describe('loadSandboxConfig', () => {
      expect(config).toEqual({
        enabled: true,
        allowedPaths: [],
-        networkAccess: false,
+        networkAccess: true,
        command: 'runsc',
        image: 'default/image',
      });
@@ -131,7 +131,7 @@ export async function loadSandboxConfig(

  let sandboxValue: boolean | string | null | undefined;
  let allowedPaths: string[] = [];
-  let networkAccess = false;
+  let networkAccess = true;
  let customImage: string | undefined;

  if (
@@ -142,7 +142,7 @@ export async function loadSandboxConfig(
    const config = sandboxOption;
    sandboxValue = config.enabled ? (config.command ?? true) : false;
    allowedPaths = config.allowedPaths ?? [];
-    networkAccess = config.networkAccess ?? false;
+    networkAccess = config.networkAccess ?? true;
    customImage = config.image;
  } else if (typeof sandboxOption !== 'object' || sandboxOption === null) {
    sandboxValue = sandboxOption;
@@ -261,7 +261,7 @@ const SETTINGS_SCHEMA = {
        requiresRestart: false,
        default: false,
        description:
-          'Enable run-event notifications for action-required prompts and session completion. Currently macOS only.',
+          'Enable run-event notifications for action-required prompts and session completion.',
        showInDialog: true,
      },
      checkpointing: {
@@ -1291,6 +1291,19 @@ const SETTINGS_SCHEMA = {
        description: 'Maximum number of directories to search for memory.',
        showInDialog: true,
      },
+      memoryBoundaryMarkers: {
+        type: 'array',
+        label: 'Memory Boundary Markers',
+        category: 'Context',
+        requiresRestart: true,
+        default: ['.git'] as string[],
+        description:
+          'File or directory names that mark the boundary for GEMINI.md discovery. ' +
+          'The upward traversal stops at the first directory containing any of these markers. ' +
+          'An empty array disables parent traversal.',
+        showInDialog: false,
+        items: { type: 'string' },
+      },
      includeDirectories: {
        type: 'array',
        label: 'Include Directories',
@@ -1445,6 +1458,21 @@ const SETTINGS_SCHEMA = {
            `,
            showInDialog: true,
          },
+          backgroundCompletionBehavior: {
+            type: 'enum',
+            label: 'Background Completion Behavior',
+            category: 'Tools',
+            requiresRestart: false,
+            default: 'silent',
+            description:
+              "Controls what happens when a background shell command finishes. 'silent' (default): quietly exits in background. 'inject': automatically returns output to agent. 'notify': shows brief message in chat.",
+            showInDialog: false,
+            options: [
+              { label: 'Silent', value: 'silent' },
+              { label: 'Inject', value: 'inject' },
+              { label: 'Notify', value: 'notify' },
+            ],
+          },
          pager: {
            type: 'string',
            label: 'Pager',
@@ -2080,6 +2108,16 @@ const SETTINGS_SCHEMA = {
          'Enable dynamic model configuration (definitions, resolutions, and chains) via settings.',
        showInDialog: false,
      },
+      useAgentProtocol: {
+        type: 'boolean',
+        label: 'Use Agent Protocol',
+        category: 'Experimental',
+        requiresRestart: true,
+        default: false,
+        description:
+          'Enable the experimental unified agent for interactive mode.',
+        showInDialog: false,
+      },
      gemmaModelRouter: {
        type: 'object',
        label: 'Gemma Model Router',
@@ -2141,6 +2179,46 @@ const SETTINGS_SCHEMA = {
          'Replace the built-in save_memory tool with a memory manager subagent that supports adding, removing, de-duplicating, and organizing memories.',
        showInDialog: true,
      },
+      agentHistoryTruncation: {
+        type: 'boolean',
+        label: 'Agent History Truncation',
+        category: 'Experimental',
+        requiresRestart: true,
+        default: false,
+        description:
+          'Enable truncation window logic for the Agent History Provider.',
+        showInDialog: true,
+      },
+      agentHistoryTruncationThreshold: {
+        type: 'number',
+        label: 'Agent History Truncation Threshold',
+        category: 'Experimental',
+        requiresRestart: true,
+        default: 30,
+        description:
+          'The maximum number of messages before history is truncated.',
+        showInDialog: true,
+      },
+      agentHistoryRetainedMessages: {
+        type: 'number',
+        label: 'Agent History Retained Messages',
+        category: 'Experimental',
+        requiresRestart: true,
+        default: 15,
+        description:
+          'The number of recent messages to retain after truncation.',
+        showInDialog: true,
+      },
+      agentHistorySummarization: {
+        type: 'boolean',
+        label: 'Agent History Summarization',
+        category: 'Experimental',
+        requiresRestart: true,
+        default: false,
+        description:
+          'Enable summarization of truncated content via a small model for the Agent History Provider.',
+        showInDialog: true,
+      },
      topicUpdateNarration: {
        type: 'boolean',
        label: 'Topic & Update Narration',
@@ -88,6 +88,8 @@ describe('Workspace-Level Policy CLI Integration', () => {
        ),
      }),
      expect.anything(),
+      undefined,
+      expect.anything(),
    );
  });

@@ -107,6 +109,8 @@ describe('Workspace-Level Policy CLI Integration', () => {
        workspacePoliciesDir: undefined,
      }),
      expect.anything(),
+      undefined,
+      expect.anything(),
    );
  });

@@ -131,6 +135,8 @@ describe('Workspace-Level Policy CLI Integration', () => {
        workspacePoliciesDir: undefined,
      }),
      expect.anything(),
+      undefined,
+      expect.anything(),
    );
  });

@@ -163,6 +169,8 @@ describe('Workspace-Level Policy CLI Integration', () => {
        ),
      }),
      expect.anything(),
+      undefined,
+      expect.anything(),
    );
  });

@@ -201,6 +209,8 @@ describe('Workspace-Level Policy CLI Integration', () => {
        ),
      }),
      expect.anything(),
+      undefined,
+      expect.anything(),
    );
  });

@@ -237,6 +247,8 @@ describe('Workspace-Level Policy CLI Integration', () => {
        ),
      }),
      expect.anything(),
+      undefined,
+      expect.anything(),
    );
  });

@@ -278,6 +290,8 @@ describe('Workspace-Level Policy CLI Integration', () => {
          workspacePoliciesDir: undefined,
        }),
        expect.anything(),
+        undefined,
+        expect.anything(),
      );
    } finally {
      // Restore for other tests
@@ -671,11 +671,6 @@ export async function main() {
      }
    }

-    // Register SessionEnd hook for graceful exit
-    registerCleanup(async () => {
-      await config.getHookSystem()?.fireSessionEndEvent(SessionEndReason.Exit);
-    });
-
    if (!input) {
      debugLogger.error(
        `No input provided via stdin. Input can be provided by piping data into gemini or using the --prompt option.`,
@@ -6,7 +6,12 @@

 import { describe, it, expect, vi, beforeEach, afterEach } from 'vitest';
 import { main } from './gemini.js';
-import { debugLogger, type Config } from '@google/gemini-cli-core';
+import {
+  debugLogger,
+  SessionEndReason,
+  type Config,
+  type HookSystem,
+} from '@google/gemini-cli-core';

 vi.mock('@google/gemini-cli-core', async (importOriginal) => {
  const actual =
@@ -197,11 +202,11 @@ describe('gemini.tsx main function cleanup', () => {
      setValue: vi.fn(),
      forScope: () => ({ settings: {}, originalSettings: {}, path: '' }),
      errors: [],
-    } as any); // eslint-disable-line @typescript-eslint/no-explicit-any
+    } as unknown as ReturnType<typeof loadSettings>);

    vi.mocked(parseArguments).mockResolvedValue({
      promptInteractive: false,
-    } as any); // eslint-disable-line @typescript-eslint/no-explicit-any
+    } as unknown as Awaited<ReturnType<typeof parseArguments>>);
    vi.mocked(loadCliConfig).mockResolvedValue({
      isInteractive: vi.fn(() => false),
      getQuestion: vi.fn(() => 'test'),
@@ -238,7 +243,8 @@ describe('gemini.tsx main function cleanup', () => {
      setTerminalBackground: vi.fn(),
      refreshAuth: vi.fn(),
      getRemoteAdminSettings: vi.fn(() => undefined),
-    } as any); // eslint-disable-line @typescript-eslint/no-explicit-any
+      getUseAlternateBuffer: vi.fn(() => false),
+    } as unknown as Config);

    await main();

@@ -248,4 +254,80 @@ describe('gemini.tsx main function cleanup', () => {
      expect.objectContaining({ message: 'Cleanup failed' }),
    );
  });
+
+  it('should register SessionEnd hook exactly once in non-interactive mode', async () => {
+    const { loadCliConfig, parseArguments } = await import(
+      './config/config.js'
+    );
+    const { registerCleanup } = await import('./utils/cleanup.js');
+
+    const mockHookSystem = {
+      fireSessionEndEvent: vi.fn().mockResolvedValue(undefined),
+      fireSessionStartEvent: vi.fn().mockResolvedValue(undefined),
+    } as unknown as HookSystem;
+
+    vi.mocked(parseArguments).mockResolvedValue({
+      promptInteractive: false,
+    } as unknown as Awaited<ReturnType<typeof parseArguments>>);
+
+    vi.mocked(loadCliConfig).mockResolvedValue(
+      buildMockConfig({
+        getHookSystem: vi.fn(() => mockHookSystem),
+      }),
+    );
+
+    vi.spyOn(process, 'exit').mockImplementation(() => undefined as never);
+
+    await main();
+
+    const registeredCallbacks = vi
+      .mocked(registerCleanup)
+      .mock.calls.map(([fn]) => fn);
+    for (const fn of registeredCallbacks) await fn();
+    expect(mockHookSystem.fireSessionEndEvent).toHaveBeenCalledTimes(1);
+    expect(mockHookSystem.fireSessionEndEvent).toHaveBeenCalledWith(
+      SessionEndReason.Exit,
+    );
+  });
+
+  function buildMockConfig(overrides: Partial<Config> = {}): Config {
+    return {
+      isInteractive: vi.fn(() => false),
+      getQuestion: vi.fn(() => 'test'),
+      getSandbox: vi.fn(() => false),
+      getDebugMode: vi.fn(() => false),
+      getPolicyEngine: vi.fn(),
+      getMessageBus: () => ({ subscribe: vi.fn() }),
+      getEnableHooks: vi.fn(() => true),
+      getHookSystem: vi.fn(() => undefined),
+      initialize: vi.fn(),
+      storage: { initialize: vi.fn().mockResolvedValue(undefined) },
+      getContentGeneratorConfig: vi.fn(),
+      getMcpClientManager: vi.fn(),
+      getIdeMode: vi.fn(() => false),
+      getAcpMode: vi.fn(() => false),
+      getScreenReader: vi.fn(() => false),
+      getGeminiMdFileCount: vi.fn(() => 0),
+      getProjectRoot: vi.fn(() => '/'),
+      getListExtensions: vi.fn(() => false),
+      getListSessions: vi.fn(() => false),
+      getDeleteSession: vi.fn(() => undefined),
+      getToolRegistry: vi.fn(),
+      getExtensions: vi.fn(() => []),
+      getModel: vi.fn(() => 'gemini-pro'),
+      getEmbeddingModel: vi.fn(() => 'embedding-001'),
+      getApprovalMode: vi.fn(() => 'default'),
+      getCoreTools: vi.fn(() => []),
+      getTelemetryEnabled: vi.fn(() => false),
+      getTelemetryLogPromptsEnabled: vi.fn(() => false),
+      getFileFilteringRespectGitIgnore: vi.fn(() => true),
+      getOutputFormat: vi.fn(() => 'text'),
+      getUsageStatisticsEnabled: vi.fn(() => false),
+      setTerminalBackground: vi.fn(),
+      refreshAuth: vi.fn(),
+      getRemoteAdminSettings: vi.fn(() => undefined),
+      getUseAlternateBuffer: vi.fn(() => false),
+      ...overrides,
+    } as unknown as Config;
+  }
 });
@@ -77,6 +77,8 @@ vi.mock('@google/gemini-cli-core', async (importOriginal) => {
    uiTelemetryService: {
      getMetrics: vi.fn(),
    },
+    LegacyAgentSession: original.LegacyAgentSession,
+    geminiPartsToContentParts: original.geminiPartsToContentParts,
    coreEvents: mockCoreEvents,
    createWorkingStdio: vi.fn(() => ({
      stdout: process.stdout,
@@ -108,6 +110,8 @@ describe('runNonInteractive', () => {
    sendMessageStream: Mock;
    resumeChat: Mock;
    getChatRecordingService: Mock;
+    getChat: Mock;
+    getCurrentSequenceModel: Mock;
  };
  const MOCK_SESSION_METRICS: SessionMetrics = {
    models: {},
@@ -163,6 +167,8 @@ describe('runNonInteractive', () => {
        recordMessageTokens: vi.fn(),
        recordToolCalls: vi.fn(),
      })),
+      getChat: vi.fn(() => ({ recordCompletedToolCalls: vi.fn() })),
+      getCurrentSequenceModel: vi.fn().mockReturnValue(null),
    };

    mockConfig = {
@@ -268,6 +274,54 @@ describe('runNonInteractive', () => {
    // so we no longer expect shutdownTelemetry to be called directly here
  });

+  it('should stream the specific stream started by send', async () => {
+    const { LegacyAgentSession } = await import('@google/gemini-cli-core');
+    const streamSpy = vi.spyOn(LegacyAgentSession.prototype, 'stream');
+    const events: ServerGeminiStreamEvent[] = [
+      { type: GeminiEventType.Content, value: 'Hello again' },
+      {
+        type: GeminiEventType.Finished,
+        value: { reason: undefined, usageMetadata: { totalTokenCount: 10 } },
+      },
+    ];
+    mockGeminiClient.sendMessageStream.mockReturnValue(
+      createStreamFromEvents(events),
+    );
+
+    await runNonInteractive({
+      config: mockConfig,
+      settings: mockSettings,
+      input: 'Test input',
+      prompt_id: 'prompt-id-stream',
+    });
+
+    expect(streamSpy).toHaveBeenCalledWith({ streamId: expect.any(String) });
+  });
+
+  it('fails fast if the session acknowledges a message send without a stream', async () => {
+    const { LegacyAgentSession } = await import('@google/gemini-cli-core');
+    const sendSpy = vi
+      .spyOn(LegacyAgentSession.prototype, 'send')
+      .mockResolvedValue({ streamId: null });
+    const streamSpy = vi.spyOn(LegacyAgentSession.prototype, 'stream');
+
+    await expect(
+      runNonInteractive({
+        config: mockConfig,
+        settings: mockSettings,
+        input: 'Test input',
+        prompt_id: 'prompt-id-null-stream',
+      }),
+    ).rejects.toThrow(
+      'LegacyAgentSession.send() unexpectedly returned no stream for a message send.',
+    );
+
+    expect(streamSpy).not.toHaveBeenCalled();
+
+    sendSpy.mockRestore();
+    streamSpy.mockRestore();
+  });
+
  it('should register activity logger when GEMINI_CLI_ACTIVITY_LOG_TARGET is set', async () => {
    vi.stubEnv('GEMINI_CLI_ACTIVITY_LOG_TARGET', '/tmp/test.jsonl');
    const events: ServerGeminiStreamEvent[] = [
@@ -558,7 +612,7 @@ describe('runNonInteractive', () => {
        input: 'Initial fail',
        prompt_id: 'prompt-id-4',
      }),
-    ).rejects.toThrow(apiError);
+    ).rejects.toThrow('API connection failed');
  });

  it('should not exit if a tool is not found, and should send error back to model', async () => {
@@ -822,6 +876,79 @@ describe('runNonInteractive', () => {
    );
  });

+  it('should keep only the final post-tool assistant text in JSON output', async () => {
+    const toolCallEvent: ServerGeminiStreamEvent = {
+      type: GeminiEventType.ToolCallRequest,
+      value: {
+        callId: 'tool-1',
+        name: 'testTool',
+        args: { arg1: 'value1' },
+        isClientInitiated: false,
+        prompt_id: 'prompt-id-json-tool-text',
+      },
+    };
+    mockSchedulerSchedule.mockResolvedValue([
+      {
+        status: CoreToolCallStatus.Success,
+        request: toolCallEvent.value,
+        tool: {} as AnyDeclarativeTool,
+        invocation: {} as AnyToolInvocation,
+        response: {
+          responseParts: [{ text: 'Tool executed successfully' }],
+          callId: 'tool-1',
+          error: undefined,
+          errorType: undefined,
+          contentLength: undefined,
+        },
+      },
+    ]);
+
+    mockGeminiClient.sendMessageStream
+      .mockReturnValueOnce(
+        createStreamFromEvents([
+          { type: GeminiEventType.Content, value: 'Let me check that...' },
+          toolCallEvent,
+          {
+            type: GeminiEventType.Finished,
+            value: { reason: undefined, usageMetadata: { totalTokenCount: 5 } },
+          },
+        ]),
+      )
+      .mockReturnValueOnce(
+        createStreamFromEvents([
+          { type: GeminiEventType.Content, value: 'Final answer' },
+          {
+            type: GeminiEventType.Finished,
+            value: { reason: undefined, usageMetadata: { totalTokenCount: 3 } },
+          },
+        ]),
+      );
+
+    vi.mocked(mockConfig.getOutputFormat).mockReturnValue(OutputFormat.JSON);
+    vi.mocked(uiTelemetryService.getMetrics).mockReturnValue(
+      MOCK_SESSION_METRICS,
+    );
+
+    await runNonInteractive({
+      config: mockConfig,
+      settings: mockSettings,
+      input: 'Use a tool',
+      prompt_id: 'prompt-id-json-tool-text',
+    });
+
+    expect(processStdoutSpy).toHaveBeenCalledWith(
+      JSON.stringify(
+        {
+          session_id: 'test-session-id',
+          response: 'Final answer',
+          stats: MOCK_SESSION_METRICS,
+        },
+        null,
+        2,
+      ),
+    );
+  });
+
  it('should write JSON output with stats for empty response commands', async () => {
    // Test the scenario where a command completes but produces no content at all
    const events: ServerGeminiStreamEvent[] = [
@@ -1068,10 +1195,11 @@ describe('runNonInteractive', () => {

    // Spy on handleCancellationError to verify it's called
    const errors = await import('./utils/errors.js');
+    const cancellationSentinel = new Error('Cancelled');
    const handleCancellationErrorSpy = vi
      .spyOn(errors, 'handleCancellationError')
      .mockImplementation(() => {
-        throw new Error('Cancelled');
+        throw cancellationSentinel;
      });

    const events: ServerGeminiStreamEvent[] = [
@@ -1087,7 +1215,7 @@ describe('runNonInteractive', () => {
            signal.addEventListener('abort', () => {
              clearTimeout(timeout);
              setTimeout(() => {
-                reject(new Error('Aborted')); // This will be caught by nonInteractiveCli and passed to handleError
+                reject(new Error('Aborted'));
              }, 300);
            });
          });
@@ -1120,20 +1248,10 @@ describe('runNonInteractive', () => {
      keypressHandler('\u0003', { ctrl: true, name: 'c' });
    }

-    // The promise should reject with 'Aborted' because our mock stream throws it,
-    // and nonInteractiveCli catches it and calls handleError, which doesn't necessarily throw.
-    // Wait, if handleError is called, we should check that.
-    // But here we want to check if Ctrl+C works.
-
-    // In our current setup, Ctrl+C aborts the signal. The stream throws 'Aborted'.
-    // nonInteractiveCli catches 'Aborted' and calls handleError.
-
-    // If we want to test that handleCancellationError is called, we need the loop to detect abortion.
-    // But our stream throws before the loop can detect it.
-
-    // Let's just check that the promise rejects with 'Aborted' for now,
-    // which proves the abortion signal reached the stream.
-    await expect(runPromise).rejects.toThrow('Aborted');
+    // The Ctrl+C path should route through handleCancellationError rather than
+    // surfacing the raw stream abort.
+    await expect(runPromise).rejects.toBe(cancellationSentinel);
+    expect(handleCancellationErrorSpy).toHaveBeenCalledTimes(1);

    expect(
      processStderrSpy.mock.calls.some(
@@ -1160,6 +1278,78 @@ describe('runNonInteractive', () => {
    // but we can also do it manually if needed.
  });

+  it('should honor cancellation that happens before session.send()', async () => {
+    const originalIsTTY = process.stdin.isTTY;
+    // eslint-disable-next-line @typescript-eslint/no-explicit-any
+    const originalSetRawMode = (process.stdin as any).setRawMode;
+
+    Object.defineProperty(process.stdin, 'isTTY', {
+      value: true,
+      configurable: true,
+    });
+    if (!originalSetRawMode) {
+      // eslint-disable-next-line @typescript-eslint/no-explicit-any
+      (process.stdin as any).setRawMode = vi.fn();
+    }
+
+    const stdinOnSpy = vi
+      .spyOn(process.stdin, 'on')
+      .mockImplementation(
+        (event: string | symbol, listener: (...args: unknown[]) => void) => {
+          if (event === 'keypress') {
+            listener('\u0003', { ctrl: true, name: 'c' });
+          }
+          return process.stdin;
+        },
+      );
+    // eslint-disable-next-line @typescript-eslint/no-explicit-any
+    vi.spyOn(process.stdin as any, 'setRawMode').mockImplementation(() => true);
+    vi.spyOn(process.stdin, 'resume').mockImplementation(() => process.stdin);
+    vi.spyOn(process.stdin, 'pause').mockImplementation(() => process.stdin);
+    vi.spyOn(process.stdin, 'removeAllListeners').mockImplementation(
+      () => process.stdin,
+    );
+
+    const errors = await import('./utils/errors.js');
+    const cancellationSentinel = new Error('Cancelled before send');
+    const handleCancellationErrorSpy = vi
+      .spyOn(errors, 'handleCancellationError')
+      .mockImplementation(() => {
+        throw cancellationSentinel;
+      });
+
+    const { LegacyAgentSession } = await import('@google/gemini-cli-core');
+    const sendSpy = vi.spyOn(LegacyAgentSession.prototype, 'send');
+
+    await expect(
+      runNonInteractive({
+        config: mockConfig,
+        settings: mockSettings,
+        input: 'Cancelled query',
+        prompt_id: 'prompt-id-pre-send-cancel',
+      }),
+    ).rejects.toBe(cancellationSentinel);
+
+    expect(handleCancellationErrorSpy).toHaveBeenCalledTimes(1);
+    expect(sendSpy).not.toHaveBeenCalled();
+    expect(stdinOnSpy).toHaveBeenCalled();
+
+    handleCancellationErrorSpy.mockRestore();
+    sendSpy.mockRestore();
+
+    Object.defineProperty(process.stdin, 'isTTY', {
+      value: originalIsTTY,
+      configurable: true,
+    });
+    if (originalSetRawMode) {
+      // eslint-disable-next-line @typescript-eslint/no-explicit-any
+      (process.stdin as any).setRawMode = originalSetRawMode;
+    } else {
+      // eslint-disable-next-line @typescript-eslint/no-explicit-any
+      delete (process.stdin as any).setRawMode;
+    }
+  });
+
  it('should throw FatalInputError if a command requires confirmation', async () => {
    const mockCommand = {
      name: 'confirm',
@@ -1329,6 +1519,9 @@ describe('runNonInteractive', () => {
        name: 'ShellTool',
        description: 'A shell tool',
        run: vi.fn(),
+        build: vi.fn().mockReturnValue({
+          getDescription: () => 'A shell tool',
+        }),
      }),
      getFunctionDeclarations: vi.fn().mockReturnValue([{ name: 'ShellTool' }]),
    } as unknown as ToolRegistry);
@@ -1777,9 +1970,7 @@ describe('runNonInteractive', () => {
        throw new Error('Recording failed');
      }),
    };
-    // @ts-expect-error - Mocking internal structure
    mockGeminiClient.getChat = vi.fn().mockReturnValue(mockChat);
-    // @ts-expect-error - Mocking internal structure
    mockGeminiClient.getCurrentSequenceModel = vi
      .fn()
      .mockReturnValue('model-1');
@@ -2000,7 +2191,6 @@ describe('runNonInteractive', () => {
      expect(processStderrSpy).toHaveBeenCalledWith(
        'Agent execution stopped: Stopped by hook\n',
      );
-      // Should exit without calling sendMessageStream again
      expect(mockGeminiClient.sendMessageStream).toHaveBeenCalledTimes(1);
    });

@@ -2031,9 +2221,9 @@ describe('runNonInteractive', () => {
      expect(processStderrSpy).toHaveBeenCalledWith(
        '[WARNING] Agent execution blocked: Blocked by hook\n',
      );
-      // sendMessageStream is called once, recursion is internal to it and transparent to the caller
-      expect(mockGeminiClient.sendMessageStream).toHaveBeenCalledTimes(1);
+      // Stream continues after blocked event — content should be output
      expect(getWrittenOutput()).toBe('Final answer\n');
+      expect(mockGeminiClient.sendMessageStream).toHaveBeenCalledTimes(1);
    });
  });

@@ -2174,6 +2364,40 @@ describe('runNonInteractive', () => {
      );
    });

+    it('should emit warning event for loop_detected in streaming JSON mode', async () => {
+      vi.mocked(mockConfig.getOutputFormat).mockReturnValue(
+        OutputFormat.STREAM_JSON,
+      );
+      vi.mocked(uiTelemetryService.getMetrics).mockReturnValue(
+        MOCK_SESSION_METRICS,
+      );
+
+      const streamEvents: ServerGeminiStreamEvent[] = [
+        { type: GeminiEventType.LoopDetected } as ServerGeminiStreamEvent,
+        { type: GeminiEventType.Content, value: 'Continuing after loop' },
+        {
+          type: GeminiEventType.Finished,
+          value: { reason: undefined, usageMetadata: { totalTokenCount: 5 } },
+        },
+      ];
+      mockGeminiClient.sendMessageStream.mockReturnValue(
+        createStreamFromEvents(streamEvents),
+      );
+
+      await runNonInteractive({
+        config: mockConfig,
+        settings: mockSettings,
+        input: 'Loop test explicit',
+        prompt_id: 'prompt-id-loop-explicit',
+      });
+
+      const output = getWrittenOutput();
+      // The STREAM_JSON output should contain an error event with warning severity
+      expect(output).toContain('"type":"error"');
+      expect(output).toContain('"severity":"warning"');
+      expect(output).toContain('Loop detected');
+    });
+
    it('should report cancelled tool calls as success in stream-json mode (legacy parity)', async () => {
      const toolCallEvent: ServerGeminiStreamEvent = {
        type: GeminiEventType.ToolCallRequest,
@@ -6,33 +6,40 @@

 import type {
  Config,
-  ToolCallRequestInfo,
  ResumedSessionData,
  UserFeedbackPayload,
+  AgentEvent,
+  ContentPart,
 } from '@google/gemini-cli-core';
 import { isSlashCommand } from './ui/utils/commandUtils.js';
 import type { LoadedSettings } from './config/settings.js';
 import {
  convertSessionToClientHistory,
-  GeminiEventType,
+  FatalError,
+  FatalAuthenticationError,
  FatalInputError,
+  FatalSandboxError,
+  FatalConfigError,
+  FatalTurnLimitedError,
+  FatalToolExecutionError,
+  FatalCancellationError,
  promptIdContext,
  OutputFormat,
  JsonFormatter,
  StreamJsonFormatter,
  JsonStreamEventType,
  uiTelemetryService,
-  debugLogger,
  coreEvents,
  CoreEvent,
  createWorkingStdio,
-  recordToolCallInteractions,
-  ToolErrorType,
  Scheduler,
  ROOT_SCHEDULER_ID,
+  LegacyAgentSession,
+  ToolErrorType,
+  geminiPartsToContentParts,
 } from '@google/gemini-cli-core';

-import type { Content, Part } from '@google/genai';
+import type { Part } from '@google/genai';
 import readline from 'node:readline';
 import stripAnsi from 'strip-ansi';

@@ -151,8 +158,6 @@ export async function runNonInteractive({
          }, 200);

          abortController.abort();
-          // Note: Don't exit here - let the abort flow through the system
-          // and trigger handleCancellationError() which will exit with proper code
        }
      };

@@ -183,6 +188,8 @@ export async function runNonInteractive({
    };

    let errorToHandle: unknown | undefined;
+    let terminalProcessExitHandled = false;
+    let abortSession = () => {};
    try {
      consolePatcher.patch();

@@ -247,9 +254,6 @@ export async function runNonInteractive({
          config,
          settings,
        );
-        // If a slash command is found and returns a prompt, use it.
-        // Otherwise, slashCommandResult falls through to the default prompt
-        // handling.
        if (slashCommandResult) {
          // eslint-disable-next-line @typescript-eslint/no-unsafe-type-assertion
          query = slashCommandResult as Part[];
@@ -267,8 +271,6 @@ export async function runNonInteractive({
          escapePastedAtSymbols: false,
        });
        if (error || !processedQuery) {
-          // An error occurred during @include processing (e.g., file not found).
-          // The error message is already logged by handleAtCommand.
          throw new FatalInputError(
            error || 'Exiting due to an error processing the @ command.',
          );
@@ -287,235 +289,334 @@ export async function runNonInteractive({
        });
      }

-      let currentMessages: Content[] = [{ role: 'user', parts: query }];
+      // Create LegacyAgentSession — owns the agentic loop
+      const session = new LegacyAgentSession({
+        client: geminiClient,
+        scheduler,
+        config,
+        promptId: prompt_id,
+      });

-      let turnCount = 0;
-      while (true) {
-        turnCount++;
-        if (
-          config.getMaxSessionTurns() >= 0 &&
-          turnCount > config.getMaxSessionTurns()
-        ) {
-          handleMaxTurnsExceededError(config);
-        }
-        const toolCallRequests: ToolCallRequestInfo[] = [];
+      // Wire Ctrl+C to session abort
+      abortSession = () => {
+        void session.abort();
+      };
+      abortController.signal.addEventListener('abort', abortSession);
+      if (abortController.signal.aborted) {
+        return handleCancellationError(config);
+      }

-        const responseStream = geminiClient.sendMessageStream(
-          currentMessages[0]?.parts || [],
-          abortController.signal,
-          prompt_id,
-          undefined,
-          false,
-          turnCount === 1 ? input : undefined,
+      // Start the agentic loop (runs in background)
+      const { streamId } = await session.send({
+        message: {
+          content: geminiPartsToContentParts(query),
+          displayContent: input,
+        },
+      });
+      if (streamId === null) {
+        throw new Error(
+          'LegacyAgentSession.send() unexpectedly returned no stream for a message send.',
        );
+      }

-        let responseText = '';
-        for await (const event of responseStream) {
-          if (abortController.signal.aborted) {
-            handleCancellationError(config);
-          }
+      const getTextContent = (parts?: ContentPart[]): string | undefined => {
+        const text = parts
+          ?.map((part) => (part.type === 'text' ? part.text : ''))
+          .join('');
+        return text ? text : undefined;
+      };

-          if (event.type === GeminiEventType.Content) {
-            const isRaw =
-              config.getRawOutput() || config.getAcceptRawOutputRisk();
-            const output = isRaw ? event.value : stripAnsi(event.value);
-            if (streamFormatter) {
-              streamFormatter.emitEvent({
-                type: JsonStreamEventType.MESSAGE,
-                timestamp: new Date().toISOString(),
-                role: 'assistant',
-                content: output,
-                delta: true,
+      const emitFinalSuccessResult = (): void => {
+        if (streamFormatter) {
+          const metrics = uiTelemetryService.getMetrics();
+          const durationMs = Date.now() - startTime;
+          streamFormatter.emitEvent({
+            type: JsonStreamEventType.RESULT,
+            timestamp: new Date().toISOString(),
+            status: 'success',
+            stats: streamFormatter.convertToStreamStats(metrics, durationMs),
+          });
+        } else if (config.getOutputFormat() === OutputFormat.JSON) {
+          const formatter = new JsonFormatter();
+          const stats = uiTelemetryService.getMetrics();
+          textOutput.write(
+            formatter.format(config.getSessionId(), responseText, stats),
+          );
+        } else {
+          textOutput.ensureTrailingNewline();
+        }
+      };
+
+      const reconstructFatalError = (event: AgentEvent<'error'>): Error => {
+        const errorMeta = event._meta;
+        const name =
+          typeof errorMeta?.['errorName'] === 'string'
+            ? errorMeta['errorName']
+            : undefined;
+
+        let errToThrow: Error;
+        switch (name) {
+          case 'FatalAuthenticationError':
+            errToThrow = new FatalAuthenticationError(event.message);
+            break;
+          case 'FatalInputError':
+            errToThrow = new FatalInputError(event.message);
+            break;
+          case 'FatalSandboxError':
+            errToThrow = new FatalSandboxError(event.message);
+            break;
+          case 'FatalConfigError':
+            errToThrow = new FatalConfigError(event.message);
+            break;
+          case 'FatalTurnLimitedError':
+            errToThrow = new FatalTurnLimitedError(event.message);
+            break;
+          case 'FatalToolExecutionError':
+            errToThrow = new FatalToolExecutionError(event.message);
+            break;
+          case 'FatalCancellationError':
+            errToThrow = new FatalCancellationError(event.message);
+            break;
+          case 'FatalError':
+            errToThrow = new FatalError(
+              event.message,
+              typeof errorMeta?.['exitCode'] === 'number'
+                ? errorMeta['exitCode']
+                : 1,
+            );
+            break;
+          default:
+            errToThrow = new Error(event.message);
+            if (name) {
+              Object.defineProperty(errToThrow, 'name', {
+                value: name,
+                enumerable: true,
              });
-            } else if (config.getOutputFormat() === OutputFormat.JSON) {
-              responseText += output;
-            } else {
-              if (event.value) {
-                textOutput.write(output);
+            }
+            break;
+        }
+
+        if (errorMeta?.['exitCode'] !== undefined) {
+          Object.defineProperty(errToThrow, 'exitCode', {
+            value: errorMeta['exitCode'],
+            enumerable: true,
+          });
+        }
+        if (errorMeta?.['code'] !== undefined) {
+          Object.defineProperty(errToThrow, 'code', {
+            value: errorMeta['code'],
+            enumerable: true,
+          });
+        }
+        if (errorMeta?.['status'] !== undefined) {
+          Object.defineProperty(errToThrow, 'status', {
+            value: errorMeta['status'],
+            enumerable: true,
+          });
+        }
+        return errToThrow;
+      };
+
+      const runTerminalExitHandler = (
+        handler: () => void | never,
+      ): void | never => {
+        terminalProcessExitHandled = true;
+        return handler();
+      };
+
+      // Consume AgentEvents for output formatting
+      let responseText = '';
+      let preToolResponseText: string | undefined;
+      let streamEnded = false;
+      for await (const event of session.stream({ streamId })) {
+        if (streamEnded) break;
+        switch (event.type) {
+          case 'message': {
+            if (event.role === 'agent') {
+              for (const part of event.content) {
+                if (part.type === 'text') {
+                  const isRaw =
+                    config.getRawOutput() || config.getAcceptRawOutputRisk();
+                  const output = isRaw ? part.text : stripAnsi(part.text);
+                  if (streamFormatter) {
+                    streamFormatter.emitEvent({
+                      type: JsonStreamEventType.MESSAGE,
+                      timestamp: new Date().toISOString(),
+                      role: 'assistant',
+                      content: output,
+                      delta: true,
+                    });
+                  } else if (config.getOutputFormat() === OutputFormat.JSON) {
+                    responseText += output;
+                  } else {
+                    if (part.text) {
+                      textOutput.write(output);
+                    }
+                  }
+                }
              }
            }
-          } else if (event.type === GeminiEventType.ToolCallRequest) {
+            break;
+          }
+          case 'tool_request': {
+            if (config.getOutputFormat() === OutputFormat.JSON) {
+              // Final JSON output should reflect the last assistant answer after
+              // any tool orchestration, not intermediate pre-tool text.
+              preToolResponseText = responseText || preToolResponseText;
+              responseText = '';
+            }
            if (streamFormatter) {
              streamFormatter.emitEvent({
                type: JsonStreamEventType.TOOL_USE,
                timestamp: new Date().toISOString(),
-                tool_name: event.value.name,
-                tool_id: event.value.callId,
-                parameters: event.value.args,
+                tool_name: event.name,
+                tool_id: event.requestId,
+                parameters: event.args,
              });
            }
-            toolCallRequests.push(event.value);
-          } else if (event.type === GeminiEventType.LoopDetected) {
-            if (streamFormatter) {
-              streamFormatter.emitEvent({
-                type: JsonStreamEventType.ERROR,
-                timestamp: new Date().toISOString(),
-                severity: 'warning',
-                message: 'Loop detected, stopping execution',
-              });
-            }
-          } else if (event.type === GeminiEventType.MaxSessionTurns) {
-            if (streamFormatter) {
-              streamFormatter.emitEvent({
-                type: JsonStreamEventType.ERROR,
-                timestamp: new Date().toISOString(),
-                severity: 'error',
-                message: 'Maximum session turns exceeded',
-              });
-            }
-          } else if (event.type === GeminiEventType.Error) {
-            throw event.value.error;
-          } else if (event.type === GeminiEventType.AgentExecutionStopped) {
-            const stopMessage = `Agent execution stopped: ${event.value.systemMessage?.trim() || event.value.reason}`;
-            if (config.getOutputFormat() === OutputFormat.TEXT) {
-              process.stderr.write(`${stopMessage}\n`);
-            }
-            // Emit final result event for streaming JSON if needed
-            if (streamFormatter) {
-              const metrics = uiTelemetryService.getMetrics();
-              const durationMs = Date.now() - startTime;
-              streamFormatter.emitEvent({
-                type: JsonStreamEventType.RESULT,
-                timestamp: new Date().toISOString(),
-                status: 'success',
-                stats: streamFormatter.convertToStreamStats(
-                  metrics,
-                  durationMs,
-                ),
-              });
-            }
-            return;
-          } else if (event.type === GeminiEventType.AgentExecutionBlocked) {
-            const blockMessage = `Agent execution blocked: ${event.value.systemMessage?.trim() || event.value.reason}`;
-            if (config.getOutputFormat() === OutputFormat.TEXT) {
-              process.stderr.write(`[WARNING] ${blockMessage}\n`);
-            }
+            break;
          }
-        }
-
-        if (toolCallRequests.length > 0) {
-          textOutput.ensureTrailingNewline();
-          const completedToolCalls = await scheduler.schedule(
-            toolCallRequests,
-            abortController.signal,
-          );
-          const toolResponseParts: Part[] = [];
-
-          for (const completedToolCall of completedToolCalls) {
-            const toolResponse = completedToolCall.response;
-            const requestInfo = completedToolCall.request;
-
+          case 'tool_response': {
+            textOutput.ensureTrailingNewline();
            if (streamFormatter) {
+              const displayText = getTextContent(event.displayContent);
+              const errorMsg = getTextContent(event.content) ?? 'Tool error';
              streamFormatter.emitEvent({
                type: JsonStreamEventType.TOOL_RESULT,
                timestamp: new Date().toISOString(),
-                tool_id: requestInfo.callId,
-                status:
-                  completedToolCall.status === 'error' ? 'error' : 'success',
-                output:
-                  typeof toolResponse.resultDisplay === 'string'
-                    ? toolResponse.resultDisplay
-                    : undefined,
-                error: toolResponse.error
+                tool_id: event.requestId,
+                status: event.isError ? 'error' : 'success',
+                output: displayText,
+                error: event.isError
                  ? {
-                      type: toolResponse.errorType || 'TOOL_EXECUTION_ERROR',
-                      message: toolResponse.error.message,
+                      type:
+                        typeof event.data?.['errorType'] === 'string'
+                          ? event.data['errorType']
+                          : 'TOOL_EXECUTION_ERROR',
+                      message: errorMsg,
                    }
                  : undefined,
              });
            }
+            if (event.isError) {
+              const displayText = getTextContent(event.displayContent);
+              const errorMsg = getTextContent(event.content) ?? 'Tool error';

-            if (toolResponse.error) {
+              if (event.data?.['errorType'] === ToolErrorType.STOP_EXECUTION) {
+                if (
+                  config.getOutputFormat() === OutputFormat.JSON &&
+                  !responseText &&
+                  preToolResponseText
+                ) {
+                  responseText = preToolResponseText;
+                }
+                const stopMessage = `Agent execution stopped: ${errorMsg}`;
+                if (config.getOutputFormat() === OutputFormat.TEXT) {
+                  process.stderr.write(`${stopMessage}\n`);
+                }
+              }
+
+              if (event.data?.['errorType'] === ToolErrorType.NO_SPACE_LEFT) {
+                terminalProcessExitHandled = true;
+                handleToolError(
+                  event.name,
+                  new Error(errorMsg),
+                  config,
+                  typeof event.data?.['errorType'] === 'string'
+                    ? event.data['errorType']
+                    : undefined,
+                  displayText,
+                );
+                return;
+              }
              handleToolError(
-                requestInfo.name,
-                toolResponse.error,
+                event.name,
+                new Error(errorMsg),
                config,
-                toolResponse.errorType || 'TOOL_EXECUTION_ERROR',
-                typeof toolResponse.resultDisplay === 'string'
-                  ? toolResponse.resultDisplay
+                typeof event.data?.['errorType'] === 'string'
+                  ? event.data['errorType']
                  : undefined,
+                displayText,
              );
            }
-
-            if (toolResponse.responseParts) {
-              toolResponseParts.push(...toolResponse.responseParts);
+            break;
+          }
+          case 'error': {
+            if (event.fatal) {
+              throw reconstructFatalError(event);
            }
-          }

-          // Record tool calls with full metadata before sending responses to Gemini
-          try {
-            const currentModel =
-              geminiClient.getCurrentSequenceModel() ?? config.getModel();
-            geminiClient
-              .getChat()
-              .recordCompletedToolCalls(currentModel, completedToolCalls);
+            const errorCode = event._meta?.['code'];

-            await recordToolCallInteractions(config, completedToolCalls);
-          } catch (error) {
-            debugLogger.error(
-              `Error recording completed tool call information: ${error}`,
-            );
-          }
-
-          // Check if any tool requested to stop execution immediately
-          const stopExecutionTool = completedToolCalls.find(
-            (tc) => tc.response.errorType === ToolErrorType.STOP_EXECUTION,
-          );
-
-          if (stopExecutionTool && stopExecutionTool.response.error) {
-            const stopMessage = `Agent execution stopped: ${stopExecutionTool.response.error.message}`;
+            if (errorCode === 'AGENT_EXECUTION_BLOCKED') {
+              if (config.getOutputFormat() === OutputFormat.TEXT) {
+                process.stderr.write(`[WARNING] ${event.message}\n`);
+              }
+              break;
+            }

+            const severity =
+              event.status === 'RESOURCE_EXHAUSTED' ? 'error' : 'warning';
            if (config.getOutputFormat() === OutputFormat.TEXT) {
-              process.stderr.write(`${stopMessage}\n`);
+              process.stderr.write(`[WARNING] ${event.message}\n`);
            }
-
-            // Emit final result event for streaming JSON
            if (streamFormatter) {
-              const metrics = uiTelemetryService.getMetrics();
-              const durationMs = Date.now() - startTime;
              streamFormatter.emitEvent({
-                type: JsonStreamEventType.RESULT,
+                type: JsonStreamEventType.ERROR,
                timestamp: new Date().toISOString(),
-                status: 'success',
-                stats: streamFormatter.convertToStreamStats(
-                  metrics,
-                  durationMs,
-                ),
+                severity,
+                message: event.message,
              });
-            } else if (config.getOutputFormat() === OutputFormat.JSON) {
-              const formatter = new JsonFormatter();
-              const stats = uiTelemetryService.getMetrics();
-              textOutput.write(
-                formatter.format(config.getSessionId(), responseText, stats),
-              );
-            } else {
-              textOutput.ensureTrailingNewline(); // Ensure a final newline
            }
-            return;
+            break;
          }
+          case 'agent_end': {
+            if (event.reason === 'aborted') {
+              runTerminalExitHandler(() => handleCancellationError(config));
+            } else if (event.reason === 'max_turns') {
+              const isConfiguredTurnLimit =
+                typeof event.data?.['maxTurns'] === 'number' ||
+                typeof event.data?.['turnCount'] === 'number';

-          currentMessages = [{ role: 'user', parts: toolResponseParts }];
-        } else {
-          // Emit final result event for streaming JSON
-          if (streamFormatter) {
-            const metrics = uiTelemetryService.getMetrics();
-            const durationMs = Date.now() - startTime;
-            streamFormatter.emitEvent({
-              type: JsonStreamEventType.RESULT,
-              timestamp: new Date().toISOString(),
-              status: 'success',
-              stats: streamFormatter.convertToStreamStats(metrics, durationMs),
-            });
-          } else if (config.getOutputFormat() === OutputFormat.JSON) {
-            const formatter = new JsonFormatter();
-            const stats = uiTelemetryService.getMetrics();
-            textOutput.write(
-              formatter.format(config.getSessionId(), responseText, stats),
-            );
-          } else {
-            textOutput.ensureTrailingNewline(); // Ensure a final newline
+              if (isConfiguredTurnLimit) {
+                runTerminalExitHandler(() =>
+                  handleMaxTurnsExceededError(config),
+                );
+              } else if (streamFormatter) {
+                streamFormatter.emitEvent({
+                  type: JsonStreamEventType.ERROR,
+                  timestamp: new Date().toISOString(),
+                  severity: 'error',
+                  message: 'Maximum session turns exceeded',
+                });
+              }
+            }
+
+            const stopMessage =
+              typeof event.data?.['message'] === 'string'
+                ? event.data['message']
+                : '';
+            if (stopMessage && config.getOutputFormat() === OutputFormat.TEXT) {
+              process.stderr.write(`Agent execution stopped: ${stopMessage}\n`);
+            }
+
+            emitFinalSuccessResult();
+            streamEnded = true;
+            break;
          }
-          return;
+          case 'initialize':
+          case 'session_update':
+          case 'agent_start':
+          case 'tool_update':
+          case 'elicitation_request':
+          case 'elicitation_response':
+          case 'usage':
+          case 'custom':
+            // Explicitly ignore these non-interactive events
+            break;
+          default:
+            event satisfies never;
+            break;
        }
      }
    } catch (error) {
@@ -523,12 +624,16 @@ export async function runNonInteractive({
    } finally {
      // Cleanup stdin cancellation before other cleanup
      cleanupStdinCancellation();
+      abortController.signal.removeEventListener('abort', abortSession);

      consolePatcher.cleanup();
      coreEvents.off(CoreEvent.UserFeedback, handleUserFeedback);
    }

    if (errorToHandle) {
+      if (terminalProcessExitHandled) {
+        throw errorToHandle;
+      }
      handleError(errorToHandle, config);
    }
  });
@@ -56,7 +56,7 @@ import { themeCommand } from '../ui/commands/themeCommand.js';
 import { toolsCommand } from '../ui/commands/toolsCommand.js';
 import { skillsCommand } from '../ui/commands/skillsCommand.js';
 import { settingsCommand } from '../ui/commands/settingsCommand.js';
-import { shellsCommand } from '../ui/commands/shellsCommand.js';
+import { tasksCommand } from '../ui/commands/tasksCommand.js';
 import { vimCommand } from '../ui/commands/vimCommand.js';
 import { setupGithubCommand } from '../ui/commands/setupGithubCommand.js';
 import { terminalSetupCommand } from '../ui/commands/terminalSetupCommand.js';
@@ -221,7 +221,7 @@ export class BuiltinCommandLoader implements ICommandLoader {
          : [skillsCommand]
        : []),
      settingsCommand,
-      shellsCommand,
+      tasksCommand,
      vimCommand,
      setupGithubCommand,
      terminalSetupCommand,
@@ -163,6 +163,7 @@ export const createMockConfig = (overrides: Partial<Config> = {}): Config =>
    getAdminSkillsEnabled: vi.fn().mockReturnValue(false),
    getDisabledSkills: vi.fn().mockReturnValue([]),
    getExperimentalJitContext: vi.fn().mockReturnValue(false),
+    getMemoryBoundaryMarkers: vi.fn().mockReturnValue(['.git']),
    getTerminalBackground: vi.fn().mockReturnValue(undefined),
    getEmbeddingModel: vi.fn().mockReturnValue('embedding-model'),
    getQuotaErrorOccurred: vi.fn().mockReturnValue(false),
@@ -0,0 +1,23 @@
+/**
+ * @license
+ * Copyright 2026 Google LLC
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+import { vi } from 'vitest';
+import type { SpinnerName } from 'cli-spinners';
+
+export function mockInkSpinner() {
+  vi.mock('ink-spinner', async () => {
+    const { Text } = await import('ink');
+    const cliSpinners = (await import('cli-spinners')).default;
+
+    return {
+      default: function MockSpinner({ type = 'dots' }: { type?: SpinnerName }) {
+        const spinner = cliSpinners[type];
+        const frame = spinner ? spinner.frames[0] : '⠋';
+        return <Text>{frame}</Text>;
+      },
+    };
+  });
+}
@@ -506,8 +506,8 @@ const baseMockUiState = {
  cleanUiDetailsVisible: false,
  allowPlanMode: true,
  activePtyId: undefined,
-  backgroundShells: new Map(),
-  backgroundShellHeight: 0,
+  backgroundTasks: new Map(),
+  backgroundTaskHeight: 0,
  quota: {
    userTier: undefined,
    stats: undefined,
@@ -568,6 +568,7 @@ const mockUIActions: UIActions = {
  handleOverageMenuChoice: vi.fn(),
  handleEmptyWalletChoice: vi.fn(),
  setQueueErrorMessage: vi.fn(),
+  addMessage: vi.fn(),
  popAllMessages: vi.fn(),
  handleApiKeySubmit: vi.fn(),
  handleApiKeyCancel: vi.fn(),
@@ -578,9 +579,9 @@ const mockUIActions: UIActions = {
  revealCleanUiDetailsTemporarily: vi.fn(),
  handleWarning: vi.fn(),
  setEmbeddedShellFocused: vi.fn(),
-  dismissBackgroundShell: vi.fn(),
-  setActiveBackgroundShellPid: vi.fn(),
-  setIsBackgroundShellListOpen: vi.fn(),
+  dismissBackgroundTask: vi.fn(),
+  setActiveBackgroundTaskPid: vi.fn(),
+  setIsBackgroundTaskListOpen: vi.fn(),
  setAuthContext: vi.fn(),
  onHintInput: vi.fn(),
  onHintBackspace: vi.fn(),
@@ -88,7 +88,7 @@ describe('App', () => {
      defaultText: 'Mock Banner Text',
      warningText: '',
    },
-    backgroundShells: new Map(),
+    backgroundTasks: new Map(),
  };

  it('should render main content and composer when not quitting', async () => {
@@ -328,13 +328,13 @@ describe('AppContainer State Management', () => {
    handleApprovalModeChange: vi.fn(),
    activePtyId: null,
    loopDetectionConfirmationRequest: null,
-    backgroundShellCount: 0,
-    isBackgroundShellVisible: false,
-    toggleBackgroundShell: vi.fn(),
-    backgroundCurrentShell: vi.fn(),
-    backgroundShells: new Map(),
-    registerBackgroundShell: vi.fn(),
-    dismissBackgroundShell: vi.fn(),
+    backgroundTaskCount: 0,
+    isBackgroundTaskVisible: false,
+    toggleBackgroundTasks: vi.fn(),
+    backgroundCurrentExecution: vi.fn(),
+    backgroundTasks: new Map(),
+    registerBackgroundTask: vi.fn(),
+    dismissBackgroundTask: vi.fn(),
  };

  beforeEach(() => {
@@ -2257,13 +2257,13 @@ describe('AppContainer State Management', () => {
      });

      it('should focus background shell on Tab when already visible (not toggle it off)', async () => {
-        const mockToggleBackgroundShell = vi.fn();
+        const mockToggleBackgroundTask = vi.fn();
        mockedUseGeminiStream.mockReturnValue({
          ...DEFAULT_GEMINI_STREAM_MOCK,
          activePtyId: null,
-          isBackgroundShellVisible: true,
-          backgroundShells: new Map([[123, { pid: 123, status: 'running' }]]),
-          toggleBackgroundShell: mockToggleBackgroundShell,
+          isBackgroundTaskVisible: true,
+          backgroundTasks: new Map([[123, { pid: 123, status: 'running' }]]),
+          toggleBackgroundTasks: mockToggleBackgroundTask,
        });

        await setupKeypressTest();
@@ -2277,7 +2277,7 @@ describe('AppContainer State Management', () => {
        // Should be focused
        expect(capturedUIState.embeddedShellFocused).toBe(true);
        // Should NOT have toggled (closed) the shell
-        expect(mockToggleBackgroundShell).not.toHaveBeenCalled();
+        expect(mockToggleBackgroundTask).not.toHaveBeenCalled();

        unmount();
      });
@@ -2285,13 +2285,13 @@ describe('AppContainer State Management', () => {

    describe('Background Shell Toggling (CTRL+B)', () => {
      it('should toggle background shell on Ctrl+B even if visible but not focused', async () => {
-        const mockToggleBackgroundShell = vi.fn();
+        const mockToggleBackgroundTask = vi.fn();
        mockedUseGeminiStream.mockReturnValue({
          ...DEFAULT_GEMINI_STREAM_MOCK,
          activePtyId: null,
-          isBackgroundShellVisible: true,
-          backgroundShells: new Map([[123, { pid: 123, status: 'running' }]]),
-          toggleBackgroundShell: mockToggleBackgroundShell,
+          isBackgroundTaskVisible: true,
+          backgroundTasks: new Map([[123, { pid: 123, status: 'running' }]]),
+          toggleBackgroundTasks: mockToggleBackgroundTask,
        });

        await setupKeypressTest();
@@ -2303,7 +2303,7 @@ describe('AppContainer State Management', () => {
        pressKey('\x02');

        // Should have toggled (closed) the shell
-        expect(mockToggleBackgroundShell).toHaveBeenCalled();
+        expect(mockToggleBackgroundTask).toHaveBeenCalled();
        // Should be unfocused
        expect(capturedUIState.embeddedShellFocused).toBe(false);

@@ -2311,28 +2311,28 @@ describe('AppContainer State Management', () => {
      });

      it('should show and focus background shell on Ctrl+B if hidden', async () => {
-        const mockToggleBackgroundShell = vi.fn();
+        const mockToggleBackgroundTask = vi.fn();
        const geminiStreamMock = {
          ...DEFAULT_GEMINI_STREAM_MOCK,
          activePtyId: null,
-          isBackgroundShellVisible: false,
-          backgroundShells: new Map([[123, { pid: 123, status: 'running' }]]),
-          toggleBackgroundShell: mockToggleBackgroundShell,
+          isBackgroundTaskVisible: false,
+          backgroundTasks: new Map([[123, { pid: 123, status: 'running' }]]),
+          toggleBackgroundTasks: mockToggleBackgroundTask,
        };
        mockedUseGeminiStream.mockReturnValue(geminiStreamMock);

        await setupKeypressTest();

        // Update the mock state when toggled to simulate real behavior
-        mockToggleBackgroundShell.mockImplementation(() => {
-          geminiStreamMock.isBackgroundShellVisible = true;
+        mockToggleBackgroundTask.mockImplementation(() => {
+          geminiStreamMock.isBackgroundTaskVisible = true;
        });

        // Press Ctrl+B
        pressKey('\x02');

        // Should have toggled (shown) the shell
-        expect(mockToggleBackgroundShell).toHaveBeenCalled();
+        expect(mockToggleBackgroundTask).toHaveBeenCalled();
        // Should be focused
        expect(capturedUIState.embeddedShellFocused).toBe(true);

@@ -82,6 +82,7 @@ import {
  buildUserSteeringHintPrompt,
  logBillingEvent,
  ApiKeyUpdatedEvent,
+  LegacyAgentProtocol,
  type InjectionSource,
 } from '@google/gemini-cli-core';
 import { validateAuthMethod } from '../config/auth.js';
@@ -110,7 +111,8 @@ import { computeTerminalTitle } from '../utils/windowTitle.js';
 import { useTextBuffer } from './components/shared/text-buffer.js';
 import { useLogger } from './hooks/useLogger.js';
 import { useGeminiStream } from './hooks/useGeminiStream.js';
-import { type BackgroundShell } from './hooks/shellCommandProcessor.js';
+import { useAgentStream } from './hooks/useAgentStream.js';
+import { type BackgroundTask } from './hooks/useExecutionLifecycle.js';
 import { useVim } from './hooks/vim.js';
 import { type LoadableSettingScope, SettingScope } from '../config/settings.js';
 import { type InitializationResult } from '../core/initializer.js';
@@ -151,7 +153,7 @@ import { useInputHistoryStore } from './hooks/useInputHistoryStore.js';
 import { useBanner } from './hooks/useBanner.js';
 import { useTerminalSetupPrompt } from './utils/terminalSetup.js';
 import { useHookDisplayState } from './hooks/useHookDisplayState.js';
-import { useBackgroundShellManager } from './hooks/useBackgroundShellManager.js';
+import { useBackgroundTaskManager } from './hooks/useBackgroundTaskManager.js';
 import {
  WARNING_PROMPT_DURATION_MS,
  QUEUE_ERROR_DISPLAY_DURATION_MS,
@@ -232,9 +234,9 @@ export const AppContainer = (props: AppContainerProps) => {
  );
  const [copyModeEnabled, setCopyModeEnabled] = useState(false);
  const [pendingRestorePrompt, setPendingRestorePrompt] = useState(false);
-  const toggleBackgroundShellRef = useRef<() => void>(() => {});
-  const isBackgroundShellVisibleRef = useRef<boolean>(false);
-  const backgroundShellsRef = useRef<Map<number, BackgroundShell>>(new Map());
+  const toggleBackgroundTasksRef = useRef<() => void>(() => {});
+  const isBackgroundTaskVisibleRef = useRef<boolean>(false);
+  const backgroundTasksRef = useRef<Map<number, BackgroundTask>>(new Map());

  const [adminSettingsChanged, setAdminSettingsChanged] = useState(false);

@@ -454,7 +456,7 @@ export const AppContainer = (props: AppContainerProps) => {

      // Kill all background shells
      await Promise.all(
-        Array.from(backgroundShellsRef.current.keys()).map((pid) =>
+        Array.from(backgroundTasksRef.current.keys()).map((pid) =>
          ShellExecutionService.kill(pid),
        ),
      );
@@ -726,7 +728,7 @@ export const AppContainer = (props: AppContainerProps) => {
  // Wrap handleDeleteSession to return a Promise for UIActions interface
  const handleDeleteSession = useCallback(
    async (session: SessionInfo): Promise<void> => {
-      handleDeleteSessionSync(session);
+      await handleDeleteSessionSync(session);
    },
    [handleDeleteSessionSync],
  );
@@ -865,7 +867,7 @@ Logging in with Google... Restarting Gemini CLI to continue.

  const { toggleVimEnabled } = useVimMode();

-  const setIsBackgroundShellListOpenRef = useRef<(open: boolean) => void>(
+  const setIsBackgroundTaskListOpenRef = useRef<(open: boolean) => void>(
    () => {},
  );
  const [shortcutsHelpVisible, setShortcutsHelpVisible] = useState(false);
@@ -900,14 +902,14 @@ Logging in with Google... Restarting Gemini CLI to continue.
      toggleDebugProfiler,
      dispatchExtensionStateUpdate,
      addConfirmUpdateExtensionRequest,
-      toggleBackgroundShell: () => {
-        toggleBackgroundShellRef.current();
-        if (!isBackgroundShellVisibleRef.current) {
+      toggleBackgroundTasks: () => {
+        toggleBackgroundTasksRef.current();
+        if (!isBackgroundTaskVisibleRef.current) {
          setEmbeddedShellFocused(true);
-          if (backgroundShellsRef.current.size > 1) {
-            setIsBackgroundShellListOpenRef.current(true);
+          if (backgroundTasksRef.current.size > 1) {
+            setIsBackgroundTaskListOpenRef.current(true);
          } else {
-            setIsBackgroundShellListOpenRef.current(false);
+            setIsBackgroundTaskListOpenRef.current(false);
          }
        }
      },
@@ -1079,7 +1081,7 @@ Logging in with Google... Restarting Gemini CLI to continue.

  useEffect(() => {
    const hintListener = (text: string, source: InjectionSource) => {
-      if (source !== 'user_steering') {
+      if (source !== 'user_steering' && source !== 'background_completion') {
        return;
      }
      pendingHintsRef.current.push(text);
@@ -1091,6 +1093,46 @@ Logging in with Google... Restarting Gemini CLI to continue.
    };
  }, [config]);

+  const streamAgent = useMemo(
+    () =>
+      config?.getExperimentalUseAgentProtocol()
+        ? new LegacyAgentProtocol({ config, getPreferredEditor })
+        : undefined,
+    [config, getPreferredEditor],
+  );
+
+  const activeStream = streamAgent
+    ? // eslint-disable-next-line react-hooks/rules-of-hooks
+      useAgentStream({
+        agent: streamAgent,
+        addItem: historyManager.addItem,
+        onCancelSubmit,
+        isShellFocused: embeddedShellFocused,
+        logger,
+      })
+    : // eslint-disable-next-line react-hooks/rules-of-hooks
+      useGeminiStream(
+        config.getGeminiClient(),
+        historyManager.history,
+        historyManager.addItem,
+        config,
+        settings,
+        setDebugMessage,
+        handleSlashCommand,
+        shellModeActive,
+        getPreferredEditor,
+        onAuthError,
+        performMemoryRefresh,
+        modelSwitchedFromQuotaError,
+        setModelSwitchedFromQuotaError,
+        onCancelSubmit,
+        setEmbeddedShellFocused,
+        terminalWidth,
+        terminalHeight,
+        embeddedShellFocused,
+        consumePendingHints,
+      );
+
  const {
    streamingState,
    submitQuery,
@@ -1103,34 +1145,14 @@ Logging in with Google... Restarting Gemini CLI to continue.
    activePtyId,
    loopDetectionConfirmationRequest,
    lastOutputTime,
-    backgroundShellCount,
-    isBackgroundShellVisible,
-    toggleBackgroundShell,
-    backgroundCurrentShell,
-    backgroundShells,
-    dismissBackgroundShell,
+    backgroundTaskCount,
+    isBackgroundTaskVisible,
+    toggleBackgroundTasks,
+    backgroundCurrentExecution,
+    backgroundTasks,
+    dismissBackgroundTask,
    retryStatus,
-  } = useGeminiStream(
-    config.getGeminiClient(),
-    historyManager.history,
-    historyManager.addItem,
-    config,
-    settings,
-    setDebugMessage,
-    handleSlashCommand,
-    shellModeActive,
-    getPreferredEditor,
-    onAuthError,
-    performMemoryRefresh,
-    modelSwitchedFromQuotaError,
-    setModelSwitchedFromQuotaError,
-    onCancelSubmit,
-    setEmbeddedShellFocused,
-    terminalWidth,
-    terminalHeight,
-    embeddedShellFocused,
-    consumePendingHints,
-  );
+  } = activeStream;

  const pendingHistoryItems = useMemo(
    () => [...pendingSlashCommandHistoryItems, ...pendingGeminiHistoryItems],
@@ -1142,27 +1164,27 @@ Logging in with Google... Restarting Gemini CLI to continue.
    [pendingHistoryItems],
  );

-  toggleBackgroundShellRef.current = toggleBackgroundShell;
-  isBackgroundShellVisibleRef.current = isBackgroundShellVisible;
-  backgroundShellsRef.current = backgroundShells;
+  toggleBackgroundTasksRef.current = toggleBackgroundTasks;
+  isBackgroundTaskVisibleRef.current = isBackgroundTaskVisible;
+  backgroundTasksRef.current = backgroundTasks;

  const {
-    activeBackgroundShellPid,
-    setIsBackgroundShellListOpen,
-    isBackgroundShellListOpen,
-    setActiveBackgroundShellPid,
-    backgroundShellHeight,
-  } = useBackgroundShellManager({
-    backgroundShells,
-    backgroundShellCount,
-    isBackgroundShellVisible,
+    activeBackgroundTaskPid,
+    setIsBackgroundTaskListOpen,
+    isBackgroundTaskListOpen,
+    setActiveBackgroundTaskPid,
+    backgroundTaskHeight,
+  } = useBackgroundTaskManager({
+    backgroundTasks,
+    backgroundTaskCount,
+    isBackgroundTaskVisible,
    activePtyId,
    embeddedShellFocused,
    setEmbeddedShellFocused,
    terminalHeight,
  });

-  setIsBackgroundShellListOpenRef.current = setIsBackgroundShellListOpen;
+  setIsBackgroundTaskListOpenRef.current = setIsBackgroundTaskListOpen;

  const lastOutputTimeRef = useRef(0);

@@ -1434,7 +1456,7 @@ Logging in with Google... Restarting Gemini CLI to continue.
  // Compute available terminal height based on stable controls measurement
  const availableTerminalHeight = Math.max(
    0,
-    terminalHeight - stableControlsHeight - backgroundShellHeight - 1,
+    terminalHeight - stableControlsHeight - backgroundTaskHeight - 1,
  );

  config.setShellExecutionConfig({
@@ -1703,7 +1725,7 @@ Logging in with Google... Restarting Gemini CLI to continue.
      if (keyMatchers[Command.QUIT](key)) {
        // If the user presses Ctrl+C, we want to cancel any ongoing requests.
        // This should happen regardless of the count.
-        cancelOngoingRequest?.();
+        void cancelOngoingRequest?.();

        handleCtrlCPress();
        return true;
@@ -1790,7 +1812,7 @@ Logging in with Google... Restarting Gemini CLI to continue.
      } else if (
        (keyMatchers[Command.FOCUS_SHELL_INPUT](key) ||
          keyMatchers[Command.UNFOCUS_BACKGROUND_SHELL_LIST](key)) &&
-        (activePtyId || (isBackgroundShellVisible && backgroundShells.size > 0))
+        (activePtyId || (isBackgroundTaskVisible && backgroundTasks.size > 0))
      ) {
        if (embeddedShellFocused) {
          const capturedTime = lastOutputTimeRef.current;
@@ -1811,12 +1833,12 @@ Logging in with Google... Restarting Gemini CLI to continue.

        const isIdle = Date.now() - lastOutputTimeRef.current >= 100;

-        if (isIdle && !activePtyId && !isBackgroundShellVisible) {
+        if (isIdle && !activePtyId && !isBackgroundTaskVisible) {
          if (tabFocusTimeoutRef.current)
            clearTimeout(tabFocusTimeoutRef.current);
-          toggleBackgroundShell();
+          toggleBackgroundTasks();
          setEmbeddedShellFocused(true);
-          if (backgroundShells.size > 1) setIsBackgroundShellListOpen(true);
+          if (backgroundTasks.size > 1) setIsBackgroundTaskListOpen(true);
          return true;
        }

@@ -1833,15 +1855,15 @@ Logging in with Google... Restarting Gemini CLI to continue.
        return false;
      } else if (keyMatchers[Command.TOGGLE_BACKGROUND_SHELL](key)) {
        if (activePtyId) {
-          backgroundCurrentShell();
+          backgroundCurrentExecution();
          // After backgrounding, we explicitly do NOT show or focus the background UI.
        } else {
-          toggleBackgroundShell();
+          toggleBackgroundTasks();
          // Toggle focus based on intent: if we were hiding, unfocus; if showing, focus.
-          if (!isBackgroundShellVisible && backgroundShells.size > 0) {
+          if (!isBackgroundTaskVisible && backgroundTasks.size > 0) {
            setEmbeddedShellFocused(true);
-            if (backgroundShells.size > 1) {
-              setIsBackgroundShellListOpen(true);
+            if (backgroundTasks.size > 1) {
+              setIsBackgroundTaskListOpen(true);
            }
          } else {
            setEmbeddedShellFocused(false);
@@ -1849,11 +1871,11 @@ Logging in with Google... Restarting Gemini CLI to continue.
        }
        return true;
      } else if (keyMatchers[Command.TOGGLE_BACKGROUND_SHELL_LIST](key)) {
-        if (backgroundShells.size > 0 && isBackgroundShellVisible) {
+        if (backgroundTasks.size > 0 && isBackgroundTaskVisible) {
          if (!embeddedShellFocused) {
            setEmbeddedShellFocused(true);
          }
-          setIsBackgroundShellListOpen(true);
+          setIsBackgroundTaskListOpen(true);
        }
        return true;
      }
@@ -1878,11 +1900,11 @@ Logging in with Google... Restarting Gemini CLI to continue.
      tabFocusTimeoutRef,
      isAlternateBuffer,
      shortcutsHelpVisible,
-      backgroundCurrentShell,
-      toggleBackgroundShell,
-      backgroundShells,
-      isBackgroundShellVisible,
-      setIsBackgroundShellListOpen,
+      backgroundCurrentExecution,
+      toggleBackgroundTasks,
+      backgroundTasks,
+      isBackgroundTaskVisible,
+      setIsBackgroundTaskListOpen,
      lastOutputTimeRef,
      showTransientMessage,
      settings.merged.general.devtools,
@@ -2055,7 +2077,7 @@ Logging in with Google... Restarting Gemini CLI to continue.
  const showStatusWit = loadingPhrases === 'witty' || loadingPhrases === 'all';

  const showLoadingIndicator =
-    (!embeddedShellFocused || isBackgroundShellVisible) &&
+    (!embeddedShellFocused || isBackgroundTaskVisible) &&
    streamingState === StreamingState.Responding &&
    !hasPendingActionRequired;

@@ -2313,8 +2335,8 @@ Logging in with Google... Restarting Gemini CLI to continue.
      isRestarting,
      extensionsUpdateState,
      activePtyId,
-      backgroundShellCount,
-      isBackgroundShellVisible,
+      backgroundTaskCount,
+      isBackgroundTaskVisible,
      embeddedShellFocused,
      showDebugProfiler,
      customDialog,
@@ -2324,10 +2346,10 @@ Logging in with Google... Restarting Gemini CLI to continue.
      bannerVisible,
      terminalBackgroundColor: config.getTerminalBackground(),
      settingsNonce,
-      backgroundShells,
-      activeBackgroundShellPid,
-      backgroundShellHeight,
-      isBackgroundShellListOpen,
+      backgroundTasks,
+      activeBackgroundTaskPid,
+      backgroundTaskHeight,
+      isBackgroundTaskListOpen,
      adminSettingsChanged,
      newAgents,
      showIsExpandableHint,
@@ -2436,8 +2458,8 @@ Logging in with Google... Restarting Gemini CLI to continue.
      currentModel,
      extensionsUpdateState,
      activePtyId,
-      backgroundShellCount,
-      isBackgroundShellVisible,
+      backgroundTaskCount,
+      isBackgroundTaskVisible,
      historyManager,
      embeddedShellFocused,
      showDebugProfiler,
@@ -2450,10 +2472,10 @@ Logging in with Google... Restarting Gemini CLI to continue.
      bannerVisible,
      config,
      settingsNonce,
-      backgroundShellHeight,
-      isBackgroundShellListOpen,
-      activeBackgroundShellPid,
-      backgroundShells,
+      backgroundTaskHeight,
+      isBackgroundTaskListOpen,
+      activeBackgroundTaskPid,
+      backgroundTasks,
      adminSettingsChanged,
      newAgents,
      showIsExpandableHint,
@@ -2502,6 +2524,7 @@ Logging in with Google... Restarting Gemini CLI to continue.
      handleResumeSession,
      handleDeleteSession,
      setQueueErrorMessage,
+      addMessage,
      popAllMessages,
      handleApiKeySubmit,
      handleApiKeyCancel,
@@ -2512,9 +2535,9 @@ Logging in with Google... Restarting Gemini CLI to continue.
      revealCleanUiDetailsTemporarily,
      handleWarning,
      setEmbeddedShellFocused,
-      dismissBackgroundShell,
-      setActiveBackgroundShellPid,
-      setIsBackgroundShellListOpen,
+      dismissBackgroundTask,
+      setActiveBackgroundTaskPid,
+      setIsBackgroundTaskListOpen,
      setAuthContext,
      onHintInput: () => {},
      onHintBackspace: () => {},
@@ -2593,6 +2616,7 @@ Logging in with Google... Restarting Gemini CLI to continue.
      handleResumeSession,
      handleDeleteSession,
      setQueueErrorMessage,
+      addMessage,
      popAllMessages,
      handleApiKeySubmit,
      handleApiKeyCancel,
@@ -2603,9 +2627,9 @@ Logging in with Google... Restarting Gemini CLI to continue.
      revealCleanUiDetailsTemporarily,
      handleWarning,
      setEmbeddedShellFocused,
-      dismissBackgroundShell,
-      setActiveBackgroundShellPid,
-      setIsBackgroundShellListOpen,
+      dismissBackgroundTask,
+      setActiveBackgroundTaskPid,
+      setIsBackgroundTaskListOpen,
      setAuthContext,
      setAccountSuspensionInfo,
      newAgents,
@@ -75,6 +75,7 @@ const listCommand: SlashCommand = {
  description: 'List saved manual conversation checkpoints',
  kind: CommandKind.BUILT_IN,
  autoExecute: true,
+  takesArgs: false,
  action: async (context): Promise<void> => {
    const chatDetails = await getSavedChatTags(context, false);

@@ -406,14 +407,24 @@ export const chatResumeSubCommands: SlashCommand[] = [
  checkpointCompatibilityCommand,
 ];

+import { parseSlashCommand } from '../../utils/commands.js';
+
 export const chatCommand: SlashCommand = {
  name: 'chat',
  description: 'Browse auto-saved conversations and manage chat checkpoints',
  kind: CommandKind.BUILT_IN,
  autoExecute: true,
-  action: async () => ({
-    type: 'dialog',
-    dialog: 'sessionBrowser',
-  }),
+  action: async (context, args) => {
+    if (args) {
+      const parsed = parseSlashCommand(`/${args}`, chatResumeSubCommands);
+      if (parsed.commandToExecute?.action) {
+        return parsed.commandToExecute.action(context, parsed.args);
+      }
+    }
+    return {
+      type: 'dialog',
+      dialog: 'sessionBrowser',
+    };
+  },
  subCommands: chatResumeSubCommands,
 };
@@ -9,6 +9,7 @@ import {
  SessionEndReason,
  SessionStartSource,
  flushTelemetry,
+  resetBrowserSession,
 } from '@google/gemini-cli-core';
 import { CommandKind, type SlashCommand } from './types.js';
 import { MessageType } from '../types.js';
@@ -43,6 +44,10 @@ export const clearCommand: SlashCommand = {

    if (geminiClient) {
      context.ui.setDebugMessage('Clearing terminal and resetting chat.');
+
+      // Close persistent browser sessions before resetting chat
+      await resetBrowserSession();
+
      // If resetChat fails, the exception will propagate and halt the command,
      // which is the correct behavior to signal a failure to the user.
      await geminiClient.resetChat();
@@ -789,6 +789,7 @@ const listExtensionsCommand: SlashCommand = {
  description: 'List active extensions',
  kind: CommandKind.BUILT_IN,
  autoExecute: true,
+  takesArgs: false,
  action: listAction,
 };

@@ -849,6 +850,7 @@ const exploreExtensionsCommand: SlashCommand = {
  description: 'Open extensions page in your browser',
  kind: CommandKind.BUILT_IN,
  autoExecute: true,
+  takesArgs: false,
  action: exploreAction,
 };

@@ -870,6 +872,8 @@ const configCommand: SlashCommand = {
  action: configAction,
 };

+import { parseSlashCommand } from '../../utils/commands.js';
+
 export function extensionsCommand(
  enableExtensionReloading?: boolean,
 ): SlashCommand {
@@ -883,20 +887,29 @@ export function extensionsCommand(
        configCommand,
      ]
    : [];
+  const subCommands = [
+    listExtensionsCommand,
+    updateExtensionsCommand,
+    exploreExtensionsCommand,
+    reloadCommand,
+    ...conditionalCommands,
+  ];
+
  return {
    name: 'extensions',
    description: 'Manage extensions',
    kind: CommandKind.BUILT_IN,
    autoExecute: false,
-    subCommands: [
-      listExtensionsCommand,
-      updateExtensionsCommand,
-      exploreExtensionsCommand,
-      reloadCommand,
-      ...conditionalCommands,
-    ],
-    action: (context, args) =>
+    subCommands,
+    action: async (context, args) => {
+      if (args) {
+        const parsed = parseSlashCommand(`/${args}`, subCommands);
+        if (parsed.commandToExecute?.action) {
+          return parsed.commandToExecute.action(context, parsed.args);
+        }
+      }
      // Default to list if no subcommand is provided
-      listExtensionsCommand.action!(context, args),
+      return listExtensionsCommand.action!(context, args);
+    },
  };
 }
@@ -17,7 +17,7 @@ import {
 } from '@google/gemini-cli-core';

 import type { CallableTool } from '@google/genai';
-import { MessageType } from '../types.js';
+import { MessageType, type HistoryItemMcpStatus } from '../types.js';

 vi.mock('@google/gemini-cli-core', async (importOriginal) => {
  const actual =
@@ -280,5 +280,41 @@ describe('mcpCommand', () => {
        }),
      );
    });
+
+    it('should filter servers by name when an argument is provided to list', async () => {
+      await mcpCommand.action!(mockContext, 'list server1');
+
+      expect(mockContext.ui.addItem).toHaveBeenCalledWith(
+        expect.objectContaining({
+          type: MessageType.MCP_STATUS,
+          servers: expect.objectContaining({
+            server1: expect.any(Object),
+          }),
+        }),
+      );
+
+      // Should NOT contain server2 or server3
+      const call = vi.mocked(mockContext.ui.addItem).mock
+        .calls[0][0] as HistoryItemMcpStatus;
+      expect(Object.keys(call.servers)).toEqual(['server1']);
+    });
+
+    it('should filter servers by name and show descriptions when an argument is provided to desc', async () => {
+      await mcpCommand.action!(mockContext, 'desc server2');
+
+      expect(mockContext.ui.addItem).toHaveBeenCalledWith(
+        expect.objectContaining({
+          type: MessageType.MCP_STATUS,
+          showDescriptions: true,
+          servers: expect.objectContaining({
+            server2: expect.any(Object),
+          }),
+        }),
+      );
+
+      const call = vi.mocked(mockContext.ui.addItem).mock
+        .calls[0][0] as HistoryItemMcpStatus;
+      expect(Object.keys(call.servers)).toEqual(['server2']);
+    });
  });
 });
@@ -31,6 +31,7 @@ import {
  canLoadServer,
 } from '../../config/mcp/mcpServerEnablement.js';
 import { loadSettings } from '../../config/settings.js';
+import { parseSlashCommand } from '../../utils/commands.js';

 const authCommand: SlashCommand = {
  name: 'auth',
@@ -177,6 +178,7 @@ const listAction = async (
  context: CommandContext,
  showDescriptions = false,
  showSchema = false,
+  serverNameFilter?: string,
 ): Promise<void | MessageActionReturn> => {
  const agentContext = context.services.agentContext;
  const config = agentContext?.config;
@@ -199,11 +201,25 @@ const listAction = async (
    };
  }

-  const mcpServers = config.getMcpClientManager()?.getMcpServers() || {};
-  const serverNames = Object.keys(mcpServers);
+  let mcpServers = config.getMcpClientManager()?.getMcpServers() || {};
  const blockedMcpServers =
    config.getMcpClientManager()?.getBlockedMcpServers() || [];

+  if (serverNameFilter) {
+    const filter = serverNameFilter.trim().toLowerCase();
+    if (filter) {
+      mcpServers = Object.fromEntries(
+        Object.entries(mcpServers).filter(
+          ([name]) =>
+            name.toLowerCase().includes(filter) ||
+            normalizeServerId(name).includes(filter),
+        ),
+      );
+    }
+  }
+
+  const serverNames = Object.keys(mcpServers);
+
  const connectingServers = serverNames.filter(
    (name) => getMCPServerStatus(name) === MCPServerStatus.CONNECTING,
  );
@@ -306,7 +322,7 @@ const listCommand: SlashCommand = {
  description: 'List configured MCP servers and tools',
  kind: CommandKind.BUILT_IN,
  autoExecute: true,
-  action: (context) => listAction(context),
+  action: (context, args) => listAction(context, false, false, args),
 };

 const descCommand: SlashCommand = {
@@ -315,7 +331,7 @@ const descCommand: SlashCommand = {
  description: 'List configured MCP servers and tools with descriptions',
  kind: CommandKind.BUILT_IN,
  autoExecute: true,
-  action: (context) => listAction(context, true),
+  action: (context, args) => listAction(context, true, false, args),
 };

 const schemaCommand: SlashCommand = {
@@ -324,7 +340,7 @@ const schemaCommand: SlashCommand = {
    'List configured MCP servers and tools with descriptions and schemas',
  kind: CommandKind.BUILT_IN,
  autoExecute: true,
-  action: (context) => listAction(context, true, true),
+  action: (context, args) => listAction(context, true, true, args),
 };

 const reloadCommand: SlashCommand = {
@@ -333,6 +349,7 @@ const reloadCommand: SlashCommand = {
  description: 'Reloads MCP servers',
  kind: CommandKind.BUILT_IN,
  autoExecute: true,
+  takesArgs: false,
  action: async (
    context: CommandContext,
  ): Promise<void | SlashCommandActionReturn> => {
@@ -530,5 +547,18 @@ export const mcpCommand: SlashCommand = {
    enableCommand,
    disableCommand,
  ],
-  action: async (context: CommandContext) => listAction(context),
+  action: async (
+    context: CommandContext,
+    args: string,
+  ): Promise<void | SlashCommandActionReturn> => {
+    if (args) {
+      const parsed = parseSlashCommand(`/${args}`, mcpCommand.subCommands!);
+      if (parsed.commandToExecute?.action) {
+        return parsed.commandToExecute.action(context, parsed.args);
+      }
+      // If no subcommand matches, treat the whole args as a filter for list
+      return listAction(context, false, false, args);
+    }
+    return listAction(context);
+  },
 };
@@ -104,6 +104,47 @@ describe('planCommand', () => {
    );
  });

+  it('should not return a submit_prompt action if arguments are empty', async () => {
+    vi.mocked(
+      mockContext.services.agentContext!.config.isPlanEnabled,
+    ).mockReturnValue(true);
+    mockContext.invocation = {
+      raw: '/plan',
+      name: 'plan',
+      args: '',
+    };
+
+    if (!planCommand.action) throw new Error('Action missing');
+    const result = await planCommand.action(mockContext, '');
+
+    expect(result).toBeUndefined();
+    expect(
+      mockContext.services.agentContext!.config.setApprovalMode,
+    ).toHaveBeenCalledWith(ApprovalMode.PLAN);
+  });
+
+  it('should return a submit_prompt action if arguments are provided', async () => {
+    vi.mocked(
+      mockContext.services.agentContext!.config.isPlanEnabled,
+    ).mockReturnValue(true);
+    mockContext.invocation = {
+      raw: '/plan implement auth',
+      name: 'plan',
+      args: 'implement auth',
+    };
+
+    if (!planCommand.action) throw new Error('Action missing');
+    const result = await planCommand.action(mockContext, 'implement auth');
+
+    expect(result).toEqual({
+      type: 'submit_prompt',
+      content: 'implement auth',
+    });
+    expect(
+      mockContext.services.agentContext!.config.setApprovalMode,
+    ).toHaveBeenCalledWith(ApprovalMode.PLAN);
+  });
+
  it('should display the approved plan from config', async () => {
    const mockPlanPath = '/mock/plans/dir/approved-plan.md';
    vi.mocked(
@@ -66,6 +66,13 @@ export const planCommand: SlashCommand = {
      coreEvents.emitFeedback('info', 'Switched to Plan Mode.');
    }

+    if (context.invocation?.args) {
+      return {
+        type: 'submit_prompt',
+        content: context.invocation.args,
+      };
+    }
+
    const approvedPlanPath = config.getApprovedPlanPath();

    if (!approvedPlanPath) {
@@ -86,12 +93,14 @@ export const planCommand: SlashCommand = {
        type: MessageType.GEMINI,
        text: partToString(content.llmContent),
      });
+      return;
    } catch (error) {
      coreEvents.emitFeedback(
        'error',
        `Failed to read approved plan at ${approvedPlanPath}: ${error}`,
        error,
      );
+      return;
    }
  },
  subCommands: [
@@ -100,6 +109,7 @@ export const planCommand: SlashCommand = {
      description: 'Copy the currently approved plan to your clipboard',
      kind: CommandKind.BUILT_IN,
      autoExecute: true,
+      takesArgs: false,
      action: copyAction,
    },
  ],
@@ -1,35 +0,0 @@
-/**
- * @license
- * Copyright 2025 Google LLC
- * SPDX-License-Identifier: Apache-2.0
- */
-
-import { describe, it, expect, vi } from 'vitest';
-import { shellsCommand } from './shellsCommand.js';
-import type { CommandContext } from './types.js';
-
-describe('shellsCommand', () => {
-  it('should call toggleBackgroundShell', async () => {
-    const toggleBackgroundShell = vi.fn();
-    const context = {
-      ui: {
-        toggleBackgroundShell,
-      },
-    } as unknown as CommandContext;
-
-    if (shellsCommand.action) {
-      await shellsCommand.action(context, '');
-    }
-
-    expect(toggleBackgroundShell).toHaveBeenCalled();
-  });
-
-  it('should have correct name and altNames', () => {
-    expect(shellsCommand.name).toBe('shells');
-    expect(shellsCommand.altNames).toContain('bashes');
-  });
-
-  it('should auto-execute', () => {
-    expect(shellsCommand.autoExecute).toBe(true);
-  });
-});
@@ -357,6 +357,8 @@ function enableCompletion(
    .map((s) => s.name);
 }

+import { parseSlashCommand } from '../../utils/commands.js';
+
 export const skillsCommand: SlashCommand = {
  name: 'skills',
  description:
@@ -402,5 +404,13 @@ export const skillsCommand: SlashCommand = {
      action: reloadAction,
    },
  ],
-  action: listAction,
+  action: async (context, args) => {
+    if (args) {
+      const parsed = parseSlashCommand(`/${args}`, skillsCommand.subCommands!);
+      if (parsed.commandToExecute?.action) {
+        return parsed.commandToExecute.action(context, parsed.args);
+      }
+    }
+    return listAction(context, args);
+  },
 };
@@ -0,0 +1,36 @@
+/**
+ * @license
+ * Copyright 2025 Google LLC
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+import { describe, it, expect, vi } from 'vitest';
+import { tasksCommand } from './tasksCommand.js';
+import type { CommandContext } from './types.js';
+
+describe('tasksCommand', () => {
+  it('should call toggleBackgroundTasks', async () => {
+    const toggleBackgroundTasks = vi.fn();
+    const context = {
+      ui: {
+        toggleBackgroundTasks,
+      },
+    } as unknown as CommandContext;
+
+    if (tasksCommand.action) {
+      await tasksCommand.action(context, '');
+    }
+
+    expect(toggleBackgroundTasks).toHaveBeenCalled();
+  });
+
+  it('should have correct name and altNames', () => {
+    expect(tasksCommand.name).toBe('tasks');
+    expect(tasksCommand.altNames).toContain('bg');
+    expect(tasksCommand.altNames).toContain('background');
+  });
+
+  it('should auto-execute', () => {
+    expect(tasksCommand.autoExecute).toBe(true);
+  });
+});
@@ -6,13 +6,13 @@

 import { CommandKind, type SlashCommand } from './types.js';

-export const shellsCommand: SlashCommand = {
-  name: 'shells',
-  altNames: ['bashes'],
+export const tasksCommand: SlashCommand = {
+  name: 'tasks',
+  altNames: ['bg', 'background'],
  kind: CommandKind.BUILT_IN,
-  description: 'Toggle background shells view',
+  description: 'Toggle background tasks view',
  autoExecute: true,
  action: async (context) => {
-    context.ui.toggleBackgroundShell();
+    context.ui.toggleBackgroundTasks();
  },
 };
@@ -90,7 +90,7 @@ export interface CommandContext {
     */
    setConfirmationRequest: (value: ConfirmationRequest) => void;
    removeComponent: () => void;
-    toggleBackgroundShell: () => void;
+    toggleBackgroundTasks: () => void;
    toggleShortcutsHelp: () => void;
  };
  // Session-specific data
@@ -240,5 +240,14 @@ export interface SlashCommand {
   */
  showCompletionLoading?: boolean;

+  /**
+   * Whether the command expects arguments.
+   * If false, and the command is a subcommand, the command parser may treat
+   * any following text as arguments for the parent command instead of this subcommand,
+   * provided the parent command has an action.
+   * Defaults to true.
+   */
+  takesArgs?: boolean;
+
  subCommands?: SlashCommand[];
 }
@@ -6,8 +6,8 @@

 import { render } from '../../test-utils/render.js';
 import { describe, it, expect, vi, beforeEach, afterEach } from 'vitest';
-import { BackgroundShellDisplay } from './BackgroundShellDisplay.js';
-import { type BackgroundShell } from '../hooks/shellCommandProcessor.js';
+import { BackgroundTaskDisplay } from './BackgroundTaskDisplay.js';
+import { type BackgroundTask } from '../hooks/useExecutionLifecycle.js';
 import { ShellExecutionService } from '@google/gemini-cli-core';
 import { act } from 'react';
 import { type Key, type KeypressHandler } from '../contexts/KeypressContext.js';
@@ -15,15 +15,15 @@ import { ScrollProvider } from '../contexts/ScrollProvider.js';
 import { Box } from 'ink';

 // Mock dependencies
-const mockDismissBackgroundShell = vi.fn();
-const mockSetActiveBackgroundShellPid = vi.fn();
-const mockSetIsBackgroundShellListOpen = vi.fn();
+const mockDismissBackgroundTask = vi.fn();
+const mockSetActiveBackgroundTaskPid = vi.fn();
+const mockSetIsBackgroundTaskListOpen = vi.fn();

 vi.mock('../contexts/UIActionsContext.js', () => ({
  useUIActions: () => ({
-    dismissBackgroundShell: mockDismissBackgroundShell,
-    setActiveBackgroundShellPid: mockSetActiveBackgroundShellPid,
-    setIsBackgroundShellListOpen: mockSetIsBackgroundShellListOpen,
+    dismissBackgroundTask: mockDismissBackgroundTask,
+    setActiveBackgroundTaskPid: mockSetActiveBackgroundTaskPid,
+    setIsBackgroundTaskListOpen: mockSetIsBackgroundTaskListOpen,
  }),
 }));

@@ -86,14 +86,14 @@ vi.mock('./shared/ScrollableList.js', () => ({
      data,
      renderItem,
    }: {
-      data: BackgroundShell[];
+      data: BackgroundTask[];
      renderItem: (props: {
-        item: BackgroundShell;
+        item: BackgroundTask;
        index: number;
      }) => React.ReactNode;
    }) => (
      <Box flexDirection="column">
-        {data.map((item: BackgroundShell, index: number) => (
+        {data.map((item: BackgroundTask, index: number) => (
          <Box key={index}>{renderItem({ item, index })}</Box>
        ))}
      </Box>
@@ -116,9 +116,9 @@ const createMockKey = (overrides: Partial<Key>): Key => ({
  ...overrides,
 });

-describe('<BackgroundShellDisplay />', () => {
-  const mockShells = new Map<number, BackgroundShell>();
-  const shell1: BackgroundShell = {
+describe('<BackgroundTaskDisplay />', () => {
+  const mockShells = new Map<number, BackgroundTask>();
+  const shell1: BackgroundTask = {
    pid: 1001,
    command: 'npm start',
    output: 'Starting server...',
@@ -126,7 +126,7 @@ describe('<BackgroundShellDisplay />', () => {
    binaryBytesReceived: 0,
    status: 'running',
  };
-  const shell2: BackgroundShell = {
+  const shell2: BackgroundTask = {
    pid: 1002,
    command: 'tail -f log.txt',
    output: 'Log entry 1',
@@ -147,7 +147,7 @@ describe('<BackgroundShellDisplay />', () => {
    const width = 80;
    const { lastFrame, unmount } = await render(
      <ScrollProvider>
-        <BackgroundShellDisplay
+        <BackgroundTaskDisplay
          shells={mockShells}
          activePid={shell1.pid}
          width={width}
@@ -167,7 +167,7 @@ describe('<BackgroundShellDisplay />', () => {
    const width = 100;
    const { lastFrame, unmount } = await render(
      <ScrollProvider>
-        <BackgroundShellDisplay
+        <BackgroundTaskDisplay
          shells={mockShells}
          activePid={shell1.pid}
          width={width}
@@ -187,7 +187,7 @@ describe('<BackgroundShellDisplay />', () => {
    const width = 80;
    const { lastFrame, unmount } = await render(
      <ScrollProvider>
-        <BackgroundShellDisplay
+        <BackgroundTaskDisplay
          shells={mockShells}
          activePid={shell1.pid}
          width={width}
@@ -207,7 +207,7 @@ describe('<BackgroundShellDisplay />', () => {
    const width = 80;
    const { rerender, unmount } = await render(
      <ScrollProvider>
-        <BackgroundShellDisplay
+        <BackgroundTaskDisplay
          shells={mockShells}
          activePid={shell1.pid}
          width={width}
@@ -227,7 +227,7 @@ describe('<BackgroundShellDisplay />', () => {

    rerender(
      <ScrollProvider>
-        <BackgroundShellDisplay
+        <BackgroundTaskDisplay
          shells={mockShells}
          activePid={shell1.pid}
          width={100}
@@ -250,7 +250,7 @@ describe('<BackgroundShellDisplay />', () => {
    const width = 80;
    const { lastFrame, unmount } = await render(
      <ScrollProvider>
-        <BackgroundShellDisplay
+        <BackgroundTaskDisplay
          shells={mockShells}
          activePid={shell1.pid}
          width={width}
@@ -270,7 +270,7 @@ describe('<BackgroundShellDisplay />', () => {
    const width = 80;
    const { unmount } = await render(
      <ScrollProvider>
-        <BackgroundShellDisplay
+        <BackgroundTaskDisplay
          shells={mockShells}
          activePid={shell1.pid}
          width={width}
@@ -287,13 +287,13 @@ describe('<BackgroundShellDisplay />', () => {
      simulateKey({ name: 'down' });
    });

-    // Simulate Ctrl+L (handled by BackgroundShellDisplay)
+    // Simulate Ctrl+L (handled by BackgroundTaskDisplay)
    await act(async () => {
      simulateKey({ name: 'l', ctrl: true });
    });

-    expect(mockSetActiveBackgroundShellPid).toHaveBeenCalledWith(shell2.pid);
-    expect(mockSetIsBackgroundShellListOpen).toHaveBeenCalledWith(false);
+    expect(mockSetActiveBackgroundTaskPid).toHaveBeenCalledWith(shell2.pid);
+    expect(mockSetIsBackgroundTaskListOpen).toHaveBeenCalledWith(false);
    unmount();
  });

@@ -301,7 +301,7 @@ describe('<BackgroundShellDisplay />', () => {
    const width = 80;
    const { unmount } = await render(
      <ScrollProvider>
-        <BackgroundShellDisplay
+        <BackgroundTaskDisplay
          shells={mockShells}
          activePid={shell1.pid}
          width={width}
@@ -325,7 +325,7 @@ describe('<BackgroundShellDisplay />', () => {
      simulateKey({ name: 'k', ctrl: true });
    });

-    expect(mockDismissBackgroundShell).toHaveBeenCalledWith(shell2.pid);
+    expect(mockDismissBackgroundTask).toHaveBeenCalledWith(shell2.pid);
    unmount();
  });

@@ -333,7 +333,7 @@ describe('<BackgroundShellDisplay />', () => {
    const width = 80;
    const { unmount } = await render(
      <ScrollProvider>
-        <BackgroundShellDisplay
+        <BackgroundTaskDisplay
          shells={mockShells}
          activePid={shell1.pid}
          width={width}
@@ -349,7 +349,7 @@ describe('<BackgroundShellDisplay />', () => {
      simulateKey({ name: 'k', ctrl: true });
    });

-    expect(mockDismissBackgroundShell).toHaveBeenCalledWith(shell1.pid);
+    expect(mockDismissBackgroundTask).toHaveBeenCalledWith(shell1.pid);
    unmount();
  });

@@ -358,7 +358,7 @@ describe('<BackgroundShellDisplay />', () => {
    const width = 80;
    const { lastFrame, unmount } = await render(
      <ScrollProvider>
-        <BackgroundShellDisplay
+        <BackgroundTaskDisplay
          shells={mockShells}
          activePid={shell2.pid}
          width={width}
@@ -375,7 +375,7 @@ describe('<BackgroundShellDisplay />', () => {
  });

  it('keeps exit code status color even when selected', async () => {
-    const exitedShell: BackgroundShell = {
+    const exitedShell: BackgroundTask = {
      pid: 1003,
      command: 'exit 0',
      output: '',
@@ -389,7 +389,7 @@ describe('<BackgroundShellDisplay />', () => {
    const width = 80;
    const { lastFrame, unmount } = await render(
      <ScrollProvider>
-        <BackgroundShellDisplay
+        <BackgroundTaskDisplay
          shells={mockShells}
          activePid={exitedShell.pid}
          width={width}
@@ -17,7 +17,7 @@ import {
  type AnsiToken,
 } from '@google/gemini-cli-core';
 import { cpLen, cpSlice, getCachedStringWidth } from '../utils/textUtils.js';
-import { type BackgroundShell } from '../hooks/shellCommandProcessor.js';
+import { type BackgroundTask } from '../hooks/useExecutionLifecycle.js';
 import { Command } from '../key/keyMatchers.js';
 import { useKeypress } from '../hooks/useKeypress.js';
 import { formatCommand } from '../key/keybindingUtils.js';
@@ -34,8 +34,8 @@ import {
 } from './shared/RadioButtonSelect.js';
 import { useKeyMatchers } from '../hooks/useKeyMatchers.js';

-interface BackgroundShellDisplayProps {
-  shells: Map<number, BackgroundShell>;
+interface BackgroundTaskDisplayProps {
+  shells: Map<number, BackgroundTask>;
  activePid: number;
  width: number;
  height: number;
@@ -61,19 +61,19 @@ const formatShellCommandForDisplay = (command: string, maxWidth: number) => {
    : commandFirstLine;
 };

-export const BackgroundShellDisplay = ({
+export const BackgroundTaskDisplay = ({
  shells,
  activePid,
  width,
  height,
  isFocused,
  isListOpenProp,
-}: BackgroundShellDisplayProps) => {
+}: BackgroundTaskDisplayProps) => {
  const keyMatchers = useKeyMatchers();
  const {
-    dismissBackgroundShell,
-    setActiveBackgroundShellPid,
-    setIsBackgroundShellListOpen,
+    dismissBackgroundTask,
+    setActiveBackgroundTaskPid,
+    setIsBackgroundTaskListOpen,
  } = useUIActions();
  const activeShell = shells.get(activePid);
  const [output, setOutput] = useState<string | AnsiOutput>(
@@ -152,13 +152,13 @@ export const BackgroundShellDisplay = ({
        // RadioButtonSelect handles Enter -> onSelect

        if (keyMatchers[Command.BACKGROUND_SHELL_ESCAPE](key)) {
-          setIsBackgroundShellListOpen(false);
+          setIsBackgroundTaskListOpen(false);
          return true;
        }

        if (keyMatchers[Command.KILL_BACKGROUND_SHELL](key)) {
          if (highlightedPid) {
-            void dismissBackgroundShell(highlightedPid);
+            void dismissBackgroundTask(highlightedPid);
            // If we killed the active one, the list might update via props
          }
          return true;
@@ -166,9 +166,9 @@ export const BackgroundShellDisplay = ({

        if (keyMatchers[Command.TOGGLE_BACKGROUND_SHELL_LIST](key)) {
          if (highlightedPid) {
-            setActiveBackgroundShellPid(highlightedPid);
+            setActiveBackgroundTaskPid(highlightedPid);
          }
-          setIsBackgroundShellListOpen(false);
+          setIsBackgroundTaskListOpen(false);
          return true;
        }
        return false;
@@ -179,12 +179,12 @@ export const BackgroundShellDisplay = ({
      }

      if (keyMatchers[Command.KILL_BACKGROUND_SHELL](key)) {
-        void dismissBackgroundShell(activeShell.pid);
+        void dismissBackgroundTask(activeShell.pid);
        return true;
      }

      if (keyMatchers[Command.TOGGLE_BACKGROUND_SHELL_LIST](key)) {
-        setIsBackgroundShellListOpen(true);
+        setIsBackgroundTaskListOpen(true);
        return true;
      }

@@ -339,8 +339,8 @@ export const BackgroundShellDisplay = ({
            items={items}
            initialIndex={initialIndex >= 0 ? initialIndex : 0}
            onSelect={(pid) => {
-              setActiveBackgroundShellPid(pid);
-              setIsBackgroundShellListOpen(false);
+              setActiveBackgroundTaskPid(pid);
+              setIsBackgroundTaskListOpen(false);
            }}
            onHighlight={(pid) => setHighlightedPid(pid)}
            isFocused={isFocused}
@@ -198,7 +198,7 @@ const createMockUIState = (overrides: Partial<UIState> = {}): UIState =>
    nightly: false,
    isTrustedFolder: true,
    activeHooks: [],
-    isBackgroundShellVisible: false,
+    isBackgroundTaskVisible: false,
    embeddedShellFocused: false,
    showIsExpandableHint: false,
    quota: {
@@ -464,7 +464,7 @@ describe('Composer', () => {
      const uiState = createMockUIState({
        streamingState: StreamingState.Responding,
        embeddedShellFocused: true,
-        isBackgroundShellVisible: true,
+        isBackgroundTaskVisible: true,
      });

      const { lastFrame } = await renderComposer(uiState);
@@ -494,7 +494,7 @@ describe('Composer', () => {
      const uiState = createMockUIState({
        streamingState: StreamingState.Responding,
        embeddedShellFocused: true,
-        isBackgroundShellVisible: false,
+        isBackgroundTaskVisible: false,
      });

      const { lastFrame } = await renderComposer(uiState);
@@ -152,6 +152,7 @@ export const Composer = ({ isFocused = true }: { isFocused?: boolean }) => {
          vimHandleInput={uiActions.vimHandleInput}
          isEmbeddedShellFocused={uiState.embeddedShellFocused}
          popAllMessages={uiActions.popAllMessages}
+          onQueueMessage={uiActions.addMessage}
          placeholder={
            vimEnabled
              ? vimMode === 'INSERT'
@@ -48,6 +48,7 @@ interface HistoryItemDisplayProps {
  isExpandable?: boolean;
  isFirstThinking?: boolean;
  isFirstAfterThinking?: boolean;
+  suppressNarration?: boolean;
 }

 export const HistoryItemDisplay: React.FC<HistoryItemDisplayProps> = ({
@@ -60,6 +61,7 @@ export const HistoryItemDisplay: React.FC<HistoryItemDisplayProps> = ({
  isExpandable,
  isFirstThinking = false,
  isFirstAfterThinking = false,
+  suppressNarration = false,
 }) => {
  const settings = useSettings();
  const inlineThinkingMode = getInlineThinkingMode(settings);
@@ -68,6 +70,17 @@ export const HistoryItemDisplay: React.FC<HistoryItemDisplayProps> = ({
  const needsTopMarginAfterThinking =
    isFirstAfterThinking && inlineThinkingMode !== 'off';

+  // If there's a topic update in this turn, we suppress the regular narration
+  // and thoughts as they are being "replaced" by the update_topic tool.
+  if (
+    suppressNarration &&
+    (itemForDisplay.type === 'thinking' ||
+      itemForDisplay.type === 'gemini' ||
+      itemForDisplay.type === 'gemini_content')
+  ) {
+    return null;
+  }
+
  return (
    <Box
      flexDirection="column"
@@ -191,6 +191,7 @@ describe('InputPrompt', () => {
    setCleanUiDetailsVisible: mockSetCleanUiDetailsVisible,
    toggleCleanUiDetailsVisible: mockToggleCleanUiDetailsVisible,
    revealCleanUiDetailsTemporarily: mockRevealCleanUiDetailsTemporarily,
+    addMessage: vi.fn(),
  };

  beforeEach(() => {
@@ -352,6 +353,8 @@ describe('InputPrompt', () => {
    vi.mocked(clipboardy.read).mockResolvedValue('');

    props = {
+      onQueueMessage: vi.fn(),
+
      buffer: mockBuffer,
      onSubmit: vi.fn(),
      userMessages: [],
@@ -1099,6 +1102,76 @@ describe('InputPrompt', () => {
    unmount();
  });

+  it('queues a message when Tab is pressed during generation', async () => {
+    props.buffer.setText('A new prompt');
+    props.streamingState = StreamingState.Responding;
+
+    const { stdin, unmount } = await renderWithProviders(
+      <InputPrompt {...props} />,
+      {
+        uiActions,
+      },
+    );
+
+    await act(async () => {
+      stdin.write('\t');
+    });
+
+    await waitFor(() => {
+      expect(props.onQueueMessage).toHaveBeenCalledWith('A new prompt');
+      expect(props.buffer.text).toBe('');
+    });
+    unmount();
+  });
+
+  it('shows an error when attempting to queue a slash command', async () => {
+    props.buffer.setText('/clear');
+    props.streamingState = StreamingState.Responding;
+
+    const { stdin, unmount } = await renderWithProviders(
+      <InputPrompt {...props} />,
+      {
+        uiActions,
+      },
+    );
+
+    await act(async () => {
+      stdin.write('\t');
+    });
+
+    await waitFor(() => {
+      expect(props.setQueueErrorMessage).toHaveBeenCalledWith(
+        'Slash commands cannot be queued',
+      );
+      expect(props.onQueueMessage).not.toHaveBeenCalled();
+    });
+    unmount();
+  });
+
+  it('shows an error when attempting to queue a shell command', async () => {
+    props.shellModeActive = true;
+    props.buffer.setText('ls');
+    props.streamingState = StreamingState.Responding;
+
+    const { stdin, unmount } = await renderWithProviders(
+      <InputPrompt {...props} />,
+      {
+        uiActions,
+      },
+    );
+
+    await act(async () => {
+      stdin.write('\t');
+    });
+
+    await waitFor(() => {
+      expect(props.setQueueErrorMessage).toHaveBeenCalledWith(
+        'Shell commands cannot be queued',
+      );
+      expect(props.onQueueMessage).not.toHaveBeenCalled();
+    });
+    unmount();
+  });
  it('should not submit on Enter when the buffer is empty or only contains whitespace', async () => {
    props.buffer.setText('   '); // Set buffer to whitespace

@@ -117,6 +117,7 @@ export interface InputPromptProps {
  setQueueErrorMessage: (message: string | null) => void;
  streamingState: StreamingState;
  popAllMessages?: () => string | undefined;
+  onQueueMessage?: (message: string) => void;
  suggestionsPosition?: 'above' | 'below';
  setBannerVisible: (visible: boolean) => void;
  copyModeEnabled?: boolean;
@@ -211,6 +212,7 @@ export const InputPrompt: React.FC<InputPromptProps> = ({
  setQueueErrorMessage,
  streamingState,
  popAllMessages,
+  onQueueMessage,
  suggestionsPosition = 'below',
  setBannerVisible,
  copyModeEnabled = false,
@@ -230,8 +232,8 @@ export const InputPrompt: React.FC<InputPromptProps> = ({
    terminalWidth,
    activePtyId,
    history,
-    backgroundShells,
-    backgroundShellHeight,
+    backgroundTasks,
+    backgroundTaskHeight,
    shortcutsHelpVisible,
  } = useUIState();
  const [suppressCompletion, setSuppressCompletion] = useState(false);
@@ -690,6 +692,7 @@ export const InputPrompt: React.FC<InputPromptProps> = ({
        streamingState === StreamingState.Responding ||
        streamingState === StreamingState.WaitingForConfirmation;

+      const isQueueMessageKey = keyMatchers[Command.QUEUE_MESSAGE](key);
      const isPlainTab =
        key.name === 'tab' && !key.shift && !key.alt && !key.ctrl && !key.cmd;
      const hasTabCompletionInteraction =
@@ -698,6 +701,29 @@ export const InputPrompt: React.FC<InputPromptProps> = ({
        reverseSearchActive ||
        commandSearchActive;

+      if (
+        isGenerating &&
+        isQueueMessageKey &&
+        !hasTabCompletionInteraction &&
+        buffer.text.trim().length > 0
+      ) {
+        const trimmedMessage = buffer.text.trim();
+        const isSlash = isSlashCommand(trimmedMessage);
+
+        if (isSlash || shellModeActive) {
+          setQueueErrorMessage(
+            `${shellModeActive ? 'Shell' : 'Slash'} commands cannot be queued`,
+          );
+        } else if (onQueueMessage) {
+          onQueueMessage(buffer.text);
+          buffer.setText('');
+          resetCompletionState();
+          resetReverseSearchCompletionState();
+        }
+        resetPlainTabPress();
+        return true;
+      }
+
      if (isPlainTab && shellModeActive) {
        resetPlainTabPress();
        if (!shouldShowSuggestions) {
@@ -1236,7 +1262,7 @@ export const InputPrompt: React.FC<InputPromptProps> = ({
      if (keyMatchers[Command.FOCUS_SHELL_INPUT](key)) {
        if (
          activePtyId ||
-          (backgroundShells.size > 0 && backgroundShellHeight > 0)
+          (backgroundTasks.size > 0 && backgroundTaskHeight > 0)
        ) {
          setEmbeddedShellFocused(true);
          return true;
@@ -1293,11 +1319,14 @@ export const InputPrompt: React.FC<InputPromptProps> = ({
      shortcutsHelpVisible,
      setShortcutsHelpVisible,
      tryLoadQueuedMessages,
+      onQueueMessage,
+      setQueueErrorMessage,
+      resetReverseSearchCompletionState,
      setBannerVisible,
      activePtyId,
      setEmbeddedShellFocused,
-      backgroundShells.size,
-      backgroundShellHeight,
+      backgroundTasks.size,
+      backgroundTaskHeight,
      streamingState,
      handleEscPress,
      registerPlainTabPress,
@@ -86,10 +86,10 @@ vi.mock('./shared/ScrollableList.js', () => ({
 }));

 import { theme } from '../semantic-colors.js';
-import { type BackgroundShell } from '../hooks/shellReducer.js';
+import { type BackgroundTask } from '../hooks/shellReducer.js';

 describe('getToolGroupBorderAppearance', () => {
-  const mockBackgroundShells = new Map<number, BackgroundShell>();
+  const mockBackgroundTasks = new Map<number, BackgroundTask>();
  const activeShellPtyId = 123;

  it('returns default empty values for non-tool_group items', () => {
@@ -99,7 +99,7 @@ describe('getToolGroupBorderAppearance', () => {
      null,
      false,
      [],
-      mockBackgroundShells,
+      mockBackgroundTasks,
    );
    expect(result).toEqual({ borderColor: '', borderDimColor: false });
  });
@@ -144,7 +144,7 @@ describe('getToolGroupBorderAppearance', () => {
      null,
      false,
      pendingItems,
-      mockBackgroundShells,
+      mockBackgroundTasks,
    );
    expect(result).toEqual({
      borderColor: theme.border.default,
@@ -173,7 +173,7 @@ describe('getToolGroupBorderAppearance', () => {
      null,
      false,
      [],
-      mockBackgroundShells,
+      mockBackgroundTasks,
    );
    expect(result).toEqual({
      borderColor: theme.border.default,
@@ -202,7 +202,7 @@ describe('getToolGroupBorderAppearance', () => {
      null,
      false,
      [],
-      mockBackgroundShells,
+      mockBackgroundTasks,
    );
    expect(result).toEqual({
      borderColor: theme.status.warning,
@@ -232,7 +232,7 @@ describe('getToolGroupBorderAppearance', () => {
      activeShellPtyId,
      false,
      [],
-      mockBackgroundShells,
+      mockBackgroundTasks,
    );
    expect(result).toEqual({
      borderColor: theme.ui.active,
@@ -262,7 +262,7 @@ describe('getToolGroupBorderAppearance', () => {
      activeShellPtyId,
      true,
      [],
-      mockBackgroundShells,
+      mockBackgroundTasks,
    );
    expect(result).toEqual({
      borderColor: theme.ui.focus,
@@ -291,7 +291,7 @@ describe('getToolGroupBorderAppearance', () => {
      activeShellPtyId,
      false,
      [],
-      mockBackgroundShells,
+      mockBackgroundTasks,
    );
    expect(result).toEqual({
      borderColor: theme.ui.active,
@@ -308,7 +308,7 @@ describe('getToolGroupBorderAppearance', () => {
      activeShellPtyId,
      true,
      [],
-      mockBackgroundShells,
+      mockBackgroundTasks,
    );
    // Since there are no tools to inspect, it falls back to empty pending, but isCurrentlyInShellTurn=true
    // so it counts as pending shell.
@@ -7,8 +7,10 @@
 import { Box, Static } from 'ink';
 import { HistoryItemDisplay } from './HistoryItemDisplay.js';
 import { useUIState } from '../contexts/UIStateContext.js';
+import { useSettings } from '../contexts/SettingsContext.js';
 import { useAppContext } from '../contexts/AppContext.js';
 import { AppHeader } from './AppHeader.js';
+
 import { useAlternateBuffer } from '../hooks/useAlternateBuffer.js';
 import {
  SCROLL_TO_ITEM_END,
@@ -19,6 +21,7 @@ import { useMemo, memo, useCallback, useEffect, useRef } from 'react';
 import { MAX_GEMINI_MESSAGE_LINES } from '../constants.js';
 import { useConfirmingTool } from '../hooks/useConfirmingTool.js';
 import { ToolConfirmationQueue } from './ToolConfirmationQueue.js';
+import { isTopicTool } from './messages/TopicMessage.js';

 const MemoizedHistoryItemDisplay = memo(HistoryItemDisplay);
 const MemoizedAppHeader = memo(AppHeader);
@@ -63,12 +66,39 @@ export const MainContent = () => {
    return -1;
  }, [uiState.history]);

+  const settings = useSettings();
+  const topicUpdateNarrationEnabled =
+    settings.merged.experimental?.topicUpdateNarration === true;
+
+  const suppressNarrationFlags = useMemo(() => {
+    const combinedHistory = [...uiState.history, ...pendingHistoryItems];
+    const flags = new Array<boolean>(combinedHistory.length).fill(false);
+
+    if (topicUpdateNarrationEnabled) {
+      let toolGroupInTurn = false;
+      for (let i = combinedHistory.length - 1; i >= 0; i--) {
+        const item = combinedHistory[i];
+        if (item.type === 'user' || item.type === 'user_shell') {
+          toolGroupInTurn = false;
+        } else if (item.type === 'tool_group') {
+          toolGroupInTurn = item.tools.some((t) => isTopicTool(t.name));
+        } else if (
+          (item.type === 'thinking' ||
+            item.type === 'gemini' ||
+            item.type === 'gemini_content') &&
+          toolGroupInTurn
+        ) {
+          flags[i] = true;
+        }
+      }
+    }
+    return flags;
+  }, [uiState.history, pendingHistoryItems, topicUpdateNarrationEnabled]);
+
  const augmentedHistory = useMemo(
    () =>
-      uiState.history.map((item, index) => {
-        const isExpandable = index > lastUserPromptIndex;
-        const prevType =
-          index > 0 ? uiState.history[index - 1]?.type : undefined;
+      uiState.history.map((item, i) => {
+        const prevType = i > 0 ? uiState.history[i - 1]?.type : undefined;
        const isFirstThinking =
          item.type === 'thinking' && prevType !== 'thinking';
        const isFirstAfterThinking =
@@ -76,18 +106,25 @@ export const MainContent = () => {

        return {
          item,
-          isExpandable,
+          isExpandable: i > lastUserPromptIndex,
          isFirstThinking,
          isFirstAfterThinking,
+          suppressNarration: suppressNarrationFlags[i] ?? false,
        };
      }),
-    [uiState.history, lastUserPromptIndex],
+    [uiState.history, lastUserPromptIndex, suppressNarrationFlags],
  );

  const historyItems = useMemo(
    () =>
      augmentedHistory.map(
-        ({ item, isExpandable, isFirstThinking, isFirstAfterThinking }) => (
+        ({
+          item,
+          isExpandable,
+          isFirstThinking,
+          isFirstAfterThinking,
+          suppressNarration,
+        }) => (
          <MemoizedHistoryItemDisplay
            terminalWidth={mainAreaWidth}
            availableTerminalHeight={
@@ -103,6 +140,7 @@ export const MainContent = () => {
            isExpandable={isExpandable}
            isFirstThinking={isFirstThinking}
            isFirstAfterThinking={isFirstAfterThinking}
+            suppressNarration={suppressNarration}
          />
        ),
      ),
@@ -138,6 +176,9 @@ export const MainContent = () => {
          const isFirstAfterThinking =
            item.type !== 'thinking' && prevType === 'thinking';

+          const suppressNarration =
+            suppressNarrationFlags[uiState.history.length + i] ?? false;
+
          return (
            <HistoryItemDisplay
              key={`pending-${i}`}
@@ -150,6 +191,7 @@ export const MainContent = () => {
              isExpandable={true}
              isFirstThinking={isFirstThinking}
              isFirstAfterThinking={isFirstAfterThinking}
+              suppressNarration={suppressNarration}
            />
          );
        })}
@@ -169,6 +211,7 @@ export const MainContent = () => {
      showConfirmationQueue,
      confirmingTool,
      uiState.history,
+      suppressNarrationFlags,
    ],
  );

@@ -176,12 +219,19 @@ export const MainContent = () => {
    () => [
      { type: 'header' as const },
      ...augmentedHistory.map(
-        ({ item, isExpandable, isFirstThinking, isFirstAfterThinking }) => ({
+        ({
+          item,
+          isExpandable,
+          isFirstThinking,
+          isFirstAfterThinking,
+          suppressNarration,
+        }) => ({
          type: 'history' as const,
          item,
          isExpandable,
          isFirstThinking,
          isFirstAfterThinking,
+          suppressNarration,
        }),
      ),
      { type: 'pending' as const },
@@ -216,6 +266,7 @@ export const MainContent = () => {
            isExpandable={item.isExpandable}
            isFirstThinking={item.isFirstThinking}
            isFirstAfterThinking={item.isFirstAfterThinking}
+            suppressNarration={item.suppressNarration}
          />
        );
      } else {
@@ -51,7 +51,7 @@ const createMockUIState = (overrides: UIStateOverrides = {}): UIState =>
    ideContextState: null,
    geminiMdFileCount: 0,
    contextFileNames: [],
-    backgroundShellCount: 0,
+    backgroundTaskCount: 0,
    buffer: { text: '' },
    history: [{ id: 1, type: 'user', text: 'test' }],
    ...overrides,
@@ -159,9 +159,9 @@ describe('StatusDisplay', () => {
    unmount();
  });

-  it('passes backgroundShellCount to ContextSummaryDisplay', async () => {
+  it('passes backgroundTaskCount to ContextSummaryDisplay', async () => {
    const uiState = createMockUIState({
-      backgroundShellCount: 3,
+      backgroundTaskCount: 3,
    });
    const { lastFrame, unmount } = await renderStatusDisplay(
      { hideContextSummary: false },
@@ -38,7 +38,7 @@ export const StatusDisplay: React.FC<StatusDisplayProps> = ({
          config.getMcpClientManager()?.getBlockedMcpServers() ?? []
        }
        skillCount={config.getSkillManager().getDisplayableSkills().length}
-        backgroundProcessCount={uiState.backgroundShellCount}
+        backgroundProcessCount={uiState.backgroundTaskCount}
      />
    );
  }
@@ -0,0 +1,145 @@
+/**
+ * @license
+ * Copyright 2026 Google LLC
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+import { renderWithProviders } from '../../test-utils/render.js';
+import { StatusRow } from './StatusRow.js';
+import { describe, it, expect, vi, beforeEach, type Mock } from 'vitest';
+import { useComposerStatus } from '../hooks/useComposerStatus.js';
+import { type UIState } from '../contexts/UIStateContext.js';
+import { type TextBuffer } from '../components/shared/text-buffer.js';
+import { type SessionStatsState } from '../contexts/SessionContext.js';
+import { type ThoughtSummary } from '../types.js';
+import { ApprovalMode } from '@google/gemini-cli-core';
+
+vi.mock('../hooks/useComposerStatus.js', () => ({
+  useComposerStatus: vi.fn(),
+}));
+
+describe('<StatusRow />', () => {
+  beforeEach(() => {
+    vi.clearAllMocks();
+  });
+
+  const defaultUiState: Partial<UIState> = {
+    currentTip: undefined,
+    thought: null,
+    elapsedTime: 0,
+    currentWittyPhrase: undefined,
+    activeHooks: [],
+    buffer: { text: '' } as unknown as TextBuffer,
+    sessionStats: { lastPromptTokenCount: 0 } as unknown as SessionStatsState,
+    shortcutsHelpVisible: false,
+    contextFileNames: [],
+    showApprovalModeIndicator: ApprovalMode.DEFAULT,
+    allowPlanMode: false,
+    shellModeActive: false,
+    renderMarkdown: true,
+    currentModel: 'gemini-3',
+  };
+
+  it('renders status and tip correctly when they both fit', async () => {
+    (useComposerStatus as Mock).mockReturnValue({
+      isInteractiveShellWaiting: false,
+      showLoadingIndicator: true,
+      showTips: true,
+      showWit: true,
+      modeContentObj: null,
+      showMinimalContext: false,
+    });
+
+    const uiState: Partial<UIState> = {
+      ...defaultUiState,
+      currentTip: 'Test Tip',
+      thought: { subject: 'Thinking...' } as unknown as ThoughtSummary,
+      elapsedTime: 5,
+      currentWittyPhrase: 'I am witty',
+    };
+
+    const { lastFrame, waitUntilReady } = await renderWithProviders(
+      <StatusRow
+        showUiDetails={false}
+        isNarrow={false}
+        terminalWidth={100}
+        hideContextSummary={false}
+        hideUiDetailsForSuggestions={false}
+        hasPendingActionRequired={false}
+      />,
+      {
+        width: 100,
+        uiState,
+      },
+    );
+
+    await waitUntilReady();
+    const output = lastFrame();
+    expect(output).toContain('Thinking...');
+    expect(output).toContain('I am witty');
+    expect(output).toContain('Tip: Test Tip');
+  });
+
+  it('renders correctly when interactive shell is waiting', async () => {
+    (useComposerStatus as Mock).mockReturnValue({
+      isInteractiveShellWaiting: true,
+      showLoadingIndicator: false,
+      showTips: false,
+      showWit: false,
+      modeContentObj: null,
+      showMinimalContext: false,
+    });
+
+    const { lastFrame, waitUntilReady } = await renderWithProviders(
+      <StatusRow
+        showUiDetails={true}
+        isNarrow={false}
+        terminalWidth={100}
+        hideContextSummary={false}
+        hideUiDetailsForSuggestions={false}
+        hasPendingActionRequired={false}
+      />,
+      {
+        width: 100,
+        uiState: defaultUiState,
+      },
+    );
+
+    await waitUntilReady();
+    expect(lastFrame()).toContain('! Shell awaiting input (Tab to focus)');
+  });
+
+  it('renders tip with absolute positioning when it fits but might collide (verification of container logic)', async () => {
+    (useComposerStatus as Mock).mockReturnValue({
+      isInteractiveShellWaiting: false,
+      showLoadingIndicator: true,
+      showTips: true,
+      showWit: true,
+      modeContentObj: null,
+      showMinimalContext: false,
+    });
+
+    const uiState: Partial<UIState> = {
+      ...defaultUiState,
+      currentTip: 'Test Tip',
+    };
+
+    const { lastFrame, waitUntilReady } = await renderWithProviders(
+      <StatusRow
+        showUiDetails={false}
+        isNarrow={false}
+        terminalWidth={100}
+        hideContextSummary={false}
+        hideUiDetailsForSuggestions={false}
+        hasPendingActionRequired={false}
+      />,
+      {
+        width: 100,
+        uiState,
+      },
+    );
+
+    await waitUntilReady();
+    expect(lastFrame()).toContain('Tip: Test Tip');
+  });
+});
@@ -179,7 +179,13 @@ export const StatusRow: React.FC<StatusRowProps> = ({
      const observer = new ResizeObserver((entries) => {
        const entry = entries[0];
        if (entry) {
-          setTipWidth(Math.round(entry.contentRect.width));
+          const width = Math.round(entry.contentRect.width);
+          // Only update if width > 0 to prevent layout feedback loops
+          // when the tip is hidden. This ensures we always use the
+          // intrinsic width for collision detection.
+          if (width > 0) {
+            setTipWidth(width);
+          }
        }
      });
      observer.observe(node);
@@ -230,6 +236,10 @@ export const StatusRow: React.FC<StatusRowProps> = ({
  const showRow1 = showUiDetails || showRow1Minimal;
  const showRow2 = showUiDetails || showRow2Minimal;

+  const onStatusResize = useCallback((width: number) => {
+    if (width > 0) setStatusWidth(width);
+  }, []);
+
  const statusNode = (
    <StatusNode
      showTips={showTips}
@@ -242,7 +252,7 @@ export const StatusRow: React.FC<StatusRowProps> = ({
      errorVerbosity={
        settings.merged.ui.errorVerbosity as 'low' | 'full' | undefined
      }
-      onResize={setStatusWidth}
+      onResize={onStatusResize}
    />
  );

@@ -322,20 +332,23 @@ export const StatusRow: React.FC<StatusRowProps> = ({

          <Box
            flexShrink={0}
-            marginLeft={LAYOUT.TIP_LEFT_MARGIN}
+            marginLeft={showTipLine ? LAYOUT.TIP_LEFT_MARGIN : 0}
            marginRight={
-              isNarrow
-                ? LAYOUT.TIP_RIGHT_MARGIN_NARROW
-                : LAYOUT.TIP_RIGHT_MARGIN_WIDE
+              showTipLine
+                ? isNarrow
+                  ? LAYOUT.TIP_RIGHT_MARGIN_NARROW
+                  : LAYOUT.TIP_RIGHT_MARGIN_WIDE
+                : 0
            }
+            position={showTipLine ? 'relative' : 'absolute'}
+            {...(showTipLine ? {} : { top: -100, left: -100 })}
          >
            {/* 
-                We always render the tip node so it can be measured by ResizeObserver,
-                but we control its visibility based on the collision detection.
+                We always render the tip node so it can be measured by ResizeObserver.
+                When hidden, we use absolute positioning so it can still be measured 
+                but doesn't affect the layout of Row 1. This prevents layout loops.
            */}
-            <Box display={showTipLine ? 'flex' : 'none'}>
-              {!isNarrow && tipContentStr && renderTipNode()}
-            </Box>
+            {!isNarrow && tipContentStr && renderTipNode()}
          </Box>
        </Box>
      )}
@@ -1,6 +1,6 @@
 // Vitest Snapshot v1, https://vitest.dev/guide/snapshot.html

-exports[`<BackgroundShellDisplay /> > highlights the focused state 1`] = `
+exports[`<BackgroundTaskDisplay /> > highlights the focused state 1`] = `
 "┌──────────────────────────────────────────────────────────────────────────────┐
 │  1: npm sta.. (PID: 1001)        Close (Ctrl+B) | Kill (Ctrl+K) | List       │
 │              (Focused)           (Ctrl+L)                                    │
@@ -10,7 +10,7 @@ exports[`<BackgroundShellDisplay /> > highlights the focused state 1`] = `
 "
 `;

-exports[`<BackgroundShellDisplay /> > keeps exit code status color even when selected 1`] = `
+exports[`<BackgroundTaskDisplay /> > keeps exit code status color even when selected 1`] = `
 "┌──────────────────────────────────────────────────────────────────────────────┐
 │  1: npm sta.. (PID: 1003)        Close (Ctrl+B) | Kill (Ctrl+K) | List       │
 │              (Focused)           (Ctrl+L)                                    │
@@ -25,7 +25,7 @@ exports[`<BackgroundShellDisplay /> > keeps exit code status color even when sel
 "
 `;

-exports[`<BackgroundShellDisplay /> > renders tabs for multiple shells 1`] = `
+exports[`<BackgroundTaskDisplay /> > renders tabs for multiple shells 1`] = `
 "┌──────────────────────────────────────────────────────────────────────────────────────────────────┐
 │  1: npm start  2: tail -f lo...  (PID: 1001)      Close (Ctrl+B) | Kill (Ctrl+K) | List (Ctrl+L) │
 │ Starting server...                                                                               │
@@ -34,7 +34,7 @@ exports[`<BackgroundShellDisplay /> > renders tabs for multiple shells 1`] = `
 "
 `;

-exports[`<BackgroundShellDisplay /> > renders the output of the active shell 1`] = `
+exports[`<BackgroundTaskDisplay /> > renders the output of the active shell 1`] = `
 "┌──────────────────────────────────────────────────────────────────────────────┐
 │  1: ...  2: ...  (PID: 1001)  Close (Ctrl+B) | Kill (Ctrl+K) | List (Ctrl+L) │
 │ Starting server...                                                           │
@@ -43,7 +43,7 @@ exports[`<BackgroundShellDisplay /> > renders the output of the active shell 1`]
 "
 `;

-exports[`<BackgroundShellDisplay /> > renders the process list when isListOpenProp is true 1`] = `
+exports[`<BackgroundTaskDisplay /> > renders the process list when isListOpenProp is true 1`] = `
 "┌──────────────────────────────────────────────────────────────────────────────┐
 │  1: npm sta.. (PID: 1001)        Close (Ctrl+B) | Kill (Ctrl+K) | List       │
 │              (Focused)           (Ctrl+L)                                    │
@@ -57,7 +57,7 @@ exports[`<BackgroundShellDisplay /> > renders the process list when isListOpenPr
 "
 `;

-exports[`<BackgroundShellDisplay /> > scrolls to active shell when list opens 1`] = `
+exports[`<BackgroundTaskDisplay /> > scrolls to active shell when list opens 1`] = `
 "┌──────────────────────────────────────────────────────────────────────────────┐
 │  1: npm sta.. (PID: 1002)        Close (Ctrl+B) | Kill (Ctrl+K) | List       │
 │              (Focused)           (Ctrl+L)                                    │
@@ -71,7 +71,7 @@
    <text x="828" y="308" fill="#afafaf" textLength="45" lengthAdjust="spacingAndGlyphs">false</text>
    <text x="891" y="308" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="0" y="325" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
-    <text x="45" y="325" fill="#afafaf" textLength="756" lengthAdjust="spacingAndGlyphs">Enable run-event notifications for action-required prompts and session completion. …</text>
+    <text x="45" y="325" fill="#afafaf" textLength="738" lengthAdjust="spacingAndGlyphs">Enable run-event notifications for action-required prompts and session completion.</text>
    <text x="891" y="325" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="0" y="342" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="891" y="342" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
@@ -71,7 +71,7 @@
    <text x="828" y="308" fill="#afafaf" textLength="45" lengthAdjust="spacingAndGlyphs">false</text>
    <text x="891" y="308" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="0" y="325" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
-    <text x="45" y="325" fill="#afafaf" textLength="756" lengthAdjust="spacingAndGlyphs">Enable run-event notifications for action-required prompts and session completion. …</text>
+    <text x="45" y="325" fill="#afafaf" textLength="738" lengthAdjust="spacingAndGlyphs">Enable run-event notifications for action-required prompts and session completion.</text>
    <text x="891" y="325" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="0" y="342" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="891" y="342" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
@@ -71,7 +71,7 @@
    <text x="828" y="308" fill="#afafaf" textLength="45" lengthAdjust="spacingAndGlyphs">false</text>
    <text x="891" y="308" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="0" y="325" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
-    <text x="45" y="325" fill="#afafaf" textLength="756" lengthAdjust="spacingAndGlyphs">Enable run-event notifications for action-required prompts and session completion. …</text>
+    <text x="45" y="325" fill="#afafaf" textLength="738" lengthAdjust="spacingAndGlyphs">Enable run-event notifications for action-required prompts and session completion.</text>
    <text x="891" y="325" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="0" y="342" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="891" y="342" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
@@ -71,7 +71,7 @@
    <text x="828" y="308" fill="#afafaf" textLength="45" lengthAdjust="spacingAndGlyphs">false</text>
    <text x="891" y="308" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="0" y="325" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
-    <text x="45" y="325" fill="#afafaf" textLength="756" lengthAdjust="spacingAndGlyphs">Enable run-event notifications for action-required prompts and session completion. …</text>
+    <text x="45" y="325" fill="#afafaf" textLength="738" lengthAdjust="spacingAndGlyphs">Enable run-event notifications for action-required prompts and session completion.</text>
    <text x="891" y="325" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="0" y="342" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="891" y="342" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
@@ -71,7 +71,7 @@
    <text x="828" y="308" fill="#afafaf" textLength="45" lengthAdjust="spacingAndGlyphs">false</text>
    <text x="891" y="308" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="0" y="325" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
-    <text x="45" y="325" fill="#afafaf" textLength="756" lengthAdjust="spacingAndGlyphs">Enable run-event notifications for action-required prompts and session completion. …</text>
+    <text x="45" y="325" fill="#afafaf" textLength="738" lengthAdjust="spacingAndGlyphs">Enable run-event notifications for action-required prompts and session completion.</text>
    <text x="891" y="325" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="0" y="342" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="891" y="342" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
@@ -60,7 +60,7 @@
    <text x="828" y="308" fill="#afafaf" textLength="45" lengthAdjust="spacingAndGlyphs">false</text>
    <text x="891" y="308" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="0" y="325" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
-    <text x="45" y="325" fill="#afafaf" textLength="756" lengthAdjust="spacingAndGlyphs">Enable run-event notifications for action-required prompts and session completion. …</text>
+    <text x="45" y="325" fill="#afafaf" textLength="738" lengthAdjust="spacingAndGlyphs">Enable run-event notifications for action-required prompts and session completion.</text>
    <text x="891" y="325" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="0" y="342" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="891" y="342" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
@@ -71,7 +71,7 @@
    <text x="828" y="308" fill="#afafaf" textLength="45" lengthAdjust="spacingAndGlyphs">false</text>
    <text x="891" y="308" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="0" y="325" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
-    <text x="45" y="325" fill="#afafaf" textLength="756" lengthAdjust="spacingAndGlyphs">Enable run-event notifications for action-required prompts and session completion. …</text>
+    <text x="45" y="325" fill="#afafaf" textLength="738" lengthAdjust="spacingAndGlyphs">Enable run-event notifications for action-required prompts and session completion.</text>
    <text x="891" y="325" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="0" y="342" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="891" y="342" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
@@ -71,7 +71,7 @@
    <text x="828" y="308" fill="#afafaf" textLength="45" lengthAdjust="spacingAndGlyphs">false</text>
    <text x="891" y="308" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="0" y="325" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
-    <text x="45" y="325" fill="#afafaf" textLength="756" lengthAdjust="spacingAndGlyphs">Enable run-event notifications for action-required prompts and session completion. …</text>
+    <text x="45" y="325" fill="#afafaf" textLength="738" lengthAdjust="spacingAndGlyphs">Enable run-event notifications for action-required prompts and session completion.</text>
    <text x="891" y="325" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="0" y="342" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="891" y="342" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
@@ -71,7 +71,7 @@
    <text x="828" y="308" fill="#afafaf" textLength="45" lengthAdjust="spacingAndGlyphs">false</text>
    <text x="891" y="308" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="0" y="325" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
-    <text x="45" y="325" fill="#afafaf" textLength="756" lengthAdjust="spacingAndGlyphs">Enable run-event notifications for action-required prompts and session completion. …</text>
+    <text x="45" y="325" fill="#afafaf" textLength="738" lengthAdjust="spacingAndGlyphs">Enable run-event notifications for action-required prompts and session completion.</text>
    <text x="891" y="325" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="0" y="342" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
    <text x="891" y="342" fill="#878787" textLength="9" lengthAdjust="spacingAndGlyphs">│</text>
@@ -20,7 +20,7 @@ exports[`SettingsDialog > Initial Rendering > should render settings list with v
 │    Enable automatic updates.                                                                     │
 │                                                                                                  │
 │    Enable Notifications                                                                   false  │
-│    Enable run-event notifications for action-required prompts and session completion. …          │
+│    Enable run-event notifications for action-required prompts and session completion.            │
 │                                                                                                  │
 │    Plan Directory                                                                     undefined  │
 │    The directory where planning artifacts are stored. If not specified, defaults t…              │
@@ -66,7 +66,7 @@ exports[`SettingsDialog > Snapshot Tests > should render 'accessibility settings
 │    Enable automatic updates.                                                                     │
 │                                                                                                  │
 │    Enable Notifications                                                                   false  │
-│    Enable run-event notifications for action-required prompts and session completion. …          │
+│    Enable run-event notifications for action-required prompts and session completion.            │
 │                                                                                                  │
 │    Plan Directory                                                                     undefined  │
 │    The directory where planning artifacts are stored. If not specified, defaults t…              │
@@ -112,7 +112,7 @@ exports[`SettingsDialog > Snapshot Tests > should render 'all boolean settings d
 │    Enable automatic updates.                                                                     │
 │                                                                                                  │
 │    Enable Notifications                                                                   false  │
-│    Enable run-event notifications for action-required prompts and session completion. …          │
+│    Enable run-event notifications for action-required prompts and session completion.            │
 │                                                                                                  │
 │    Plan Directory                                                                     undefined  │
 │    The directory where planning artifacts are stored. If not specified, defaults t…              │
@@ -158,7 +158,7 @@ exports[`SettingsDialog > Snapshot Tests > should render 'default state' correct
 │    Enable automatic updates.                                                                     │
 │                                                                                                  │
 │    Enable Notifications                                                                   false  │
-│    Enable run-event notifications for action-required prompts and session completion. …          │
+│    Enable run-event notifications for action-required prompts and session completion.            │
 │                                                                                                  │
 │    Plan Directory                                                                     undefined  │
 │    The directory where planning artifacts are stored. If not specified, defaults t…              │
@@ -204,7 +204,7 @@ exports[`SettingsDialog > Snapshot Tests > should render 'file filtering setting
 │    Enable automatic updates.                                                                     │
 │                                                                                                  │
 │    Enable Notifications                                                                   false  │
-│    Enable run-event notifications for action-required prompts and session completion. …          │
+│    Enable run-event notifications for action-required prompts and session completion.            │
 │                                                                                                  │
 │    Plan Directory                                                                     undefined  │
 │    The directory where planning artifacts are stored. If not specified, defaults t…              │
@@ -250,7 +250,7 @@ exports[`SettingsDialog > Snapshot Tests > should render 'focused on scope selec
 │    Enable automatic updates.                                                                     │
 │                                                                                                  │
 │    Enable Notifications                                                                   false  │
-│    Enable run-event notifications for action-required prompts and session completion. …          │
+│    Enable run-event notifications for action-required prompts and session completion.            │
 │                                                                                                  │
 │    Plan Directory                                                                     undefined  │
 │    The directory where planning artifacts are stored. If not specified, defaults t…              │
@@ -296,7 +296,7 @@ exports[`SettingsDialog > Snapshot Tests > should render 'mixed boolean and numb
 │    Enable automatic updates.                                                                     │
 │                                                                                                  │
 │    Enable Notifications                                                                   false  │
-│    Enable run-event notifications for action-required prompts and session completion. …          │
+│    Enable run-event notifications for action-required prompts and session completion.            │
 │                                                                                                  │
 │    Plan Directory                                                                     undefined  │
 │    The directory where planning artifacts are stored. If not specified, defaults t…              │
@@ -342,7 +342,7 @@ exports[`SettingsDialog > Snapshot Tests > should render 'tools and security set
 │    Enable automatic updates.                                                                     │
 │                                                                                                  │
 │    Enable Notifications                                                                   false  │
-│    Enable run-event notifications for action-required prompts and session completion. …          │
+│    Enable run-event notifications for action-required prompts and session completion.            │
 │                                                                                                  │
 │    Plan Directory                                                                     undefined  │
 │    The directory where planning artifacts are stored. If not specified, defaults t…              │
@@ -388,7 +388,7 @@ exports[`SettingsDialog > Snapshot Tests > should render 'various boolean settin
 │    Enable automatic updates.                                                                     │
 │                                                                                                  │
 │    Enable Notifications                                                                   false  │
-│    Enable run-event notifications for action-required prompts and session completion. …          │
+│    Enable run-event notifications for action-required prompts and session completion.            │
 │                                                                                                  │
 │    Plan Directory                                                                     undefined  │
 │    The directory where planning artifacts are stored. If not specified, defaults t…              │
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
Michael Bleigh	d9e2c87131	Merge branch 'pr4-cli' into mb/agent-interactive	2026-03-30 12:59:33 -07:00
Michael Bleigh	3b0906f319	Merge branch 'main' into pr4-cli	2026-03-30 12:39:30 -07:00
Michael Bleigh	cd7d91f35e	refactor(cli): remove config dependency from useAgentStream	2026-03-30 12:35:31 -07:00
Michael Bleigh	6020551d1f	Merge branch 'pr4-cli' of github.com:google-gemini/gemini-cli into mb/agent-interactive	2026-03-30 09:48:31 -07:00
Alex Stephen	9cf410478c	Text can be added after /plan command (#22833 )	2026-03-30 14:31:20 +00:00
Spencer	a255529c6b	fix(a2a-server): prioritize ADC before evaluating headless constraints for auth initialization (#23614 )	2026-03-30 12:26:15 +00:00
Abhi	d9d2ce36f2	test(evals): add comprehensive subagent delegation evaluations (#24132 )	2026-03-29 23:13:50 +00:00
Adam Weidman	8e9961a791	fixes	2026-03-29 17:43:07 -04:00
Tommaso Sciortino	da8c841ef4	fix: shellcheck warnings in scripts (#24035 )	2026-03-29 02:47:05 +00:00
Christian Gunderman	b7c86b5497	UX for topic narration tool (#24079 )	2026-03-28 21:33:38 +00:00
Adam Weidman	3eebb75b7a	feat(core): agnostic background task UI with CompletionBehavior (#22740 ) Co-authored-by: mkorwel <matt.korwel@gmail.com>	2026-03-28 21:27:51 +00:00
Christian Gunderman	07ab16dbbe	feat(cli): support 'tab to queue' for messages while generating (#24052 )	2026-03-28 01:31:11 +00:00
Abhijit Balaji	afc1d50c20	feat(core): implement tool-based topic grouping (Chapters) (#23150 ) Co-authored-by: Christian Gunderman <gundermanc@google.com>	2026-03-28 01:28:25 +00:00
Gal Zahavi	ae123c547c	fix(sandbox): implement Windows Mandatory Integrity Control for GeminiSandbox (#24057 )	2026-03-28 00:14:35 +00:00
Keith Guerin	c2705e8332	fix(cli): resolve layout contention and flashing loop in StatusRow (#24065 )	2026-03-28 00:06:07 +00:00
Sam Roberts	9574855435	Re-word intro to Gemini 3 page. (#24069 )	2026-03-28 00:01:22 +00:00
krishdef7	bf6dae4690	fix(hooks): prevent SessionEnd from firing twice in non-interactive mode (#22139 ) Co-authored-by: Tommaso Sciortino <sciortino@gmail.com>	2026-03-27 23:39:48 +00:00
Ivan Porty	b5529c2475	fix(core): resolve ACP Operation Aborted Errors in grep_search (#23821 ) Co-authored-by: Sri Pasumarthi <sripas@google.com>	2026-03-27 23:12:34 +00:00
Michael Bleigh	e046cbd8e0	fix: resolve rebase compilation and test failures	2026-03-27 16:05:06 -07:00
Michael Bleigh	b810e57ef1	feat: add experimental useAgentProtocol flag	2026-03-27 15:42:12 -07:00
Michael Bleigh	ecc9e50a1f	fix: resolve typescript lint errors and test failures - Remove unnecessary `any` casts and unsafe type assertions in `useAgentStream.ts`. - Introduce `MinimalTrackedToolCall` to safely type mock tool calls for inactivity monitors. - Fix arrow-body-style lint errors in `AppContainer.tsx` and `useAgentStream.ts`. - Update `nonInteractiveCli.test.ts` to include a required `build` method in mock tools to prevent TypeErrors during stream initialization. - Remove redundant non-null assertion in `legacy-agent-session.ts`.	2026-03-27 15:41:39 -07:00
joshualitt	fd321abd3d	feat(core): Land `AgentHistoryProvider`. (#23978 )	2026-03-27 15:41:39 -07:00
Sri Pasumarthi	2bb161b36f	fix(acp): prevent crash on empty response in ACP mode (#23952 )	2026-03-27 15:35:58 -07:00
Jacob Richman	565851f2d5	Increase memory limited for eslint. (#24022 )	2026-03-27 15:35:58 -07:00
Michael Bleigh	593c33f927	refactor(cli): simplify useAgentStream state to use IndividualToolCallDisplay This commit refactors the `useAgentStream` hook to track its internal state using the lightweight `IndividualToolCallDisplay` interface instead of the heavyweight `TrackedToolCall`. By mapping `AgentEvent` payloads directly to `IndividualToolCallDisplay`, we completely bypass the need for `DummyTool` re-hydration and the `mapToDisplay` adapter. This removes a redundant data bridging layer and properly aligns the UI state with the flattened data provided by the `AgentProtocol` in `legacyState`.	2026-03-27 15:35:58 -07:00
Michael Bleigh	f85299717a	refactor: allow GEMINI_CLI_USE_AGENT_PROTOCOL env to trigger agent use	2026-03-27 15:35:58 -07:00
Michael Bleigh	0afe5117a4	fix(cli,core): resolve lint and type errors in agent stream and core types	2026-03-27 15:35:54 -07:00
Tommaso Sciortino	9e74a7ec18	feat(core): implement windows sandbox expansion and denial detection (#24027 )	2026-03-27 22:35:01 +00:00
Michael Bleigh	f853d2f9da	fix: explicitly fail pending tool calls when approval is required This commit ensures that when a tool call enters the 'awaiting_approval' state (which is not yet supported in the new protocol), we not only emit a fatal error but also emit a 'tool_response' event with 'isError: true'. This allows the UI to correctly transition the tool call from a pending (scheduled/executing) state to an error state, providing clear visual feedback to the user and preventing the UI from appearing to hang.	2026-03-27 15:02:25 -07:00
Michael Bleigh	0f37dd1d78	fix: error out on tool approvals instead of hanging This commit ensures that the interactive agent loop does not hang when a tool requires manual approval (since confirmation dialogs are not yet implemented for the new protocol). Instead of waiting indefinitely, 'LegacyAgentSession' now detects the 'awaiting_approval' status, emits a descriptive error event, and aborts the session. Key changes: - 'LegacyAgentSession' now emits status updates in 'tool_update' events. - 'LegacyAgentSession' aborts and errors on 'awaiting_approval' status. - 'useAgentStream' now correctly tracks and displays tool status transitions and intermediate progress via the 'AgentEvent' stream.	2026-03-27 15:02:25 -07:00
Michael Bleigh	87e9491e78	fix: ensure correct message ordering in agent stream This commit fixes a bug where model text preceding a tool call would be displayed after the tool call in the conversation history. We now explicitly flush any pending agent text to the conversation history (via 'addItem') immediately before a tool request is tracked. This ensures the model's intent is rendered above the execution result.	2026-03-27 15:02:25 -07:00
Michael Bleigh	2a8918c72f	feat: achieve rich tool call display parity for agent stream This change ensures that tool execution in the Agent Protocol experiment reaches visual parity with the legacy implementation by passing rich metadata and objects (like FileDiff) through the AgentEvent stream. Key changes: - Core 'LegacyAgentSession' now builds temporary invocations for tool requests to capture rich, argument-aware descriptions (e.g. 'Writing to poem.md'). - Core 'LegacyAgentSession' now attaches the raw 'resultDisplay' object and 'outputFile' path to the 'tool_response' event metadata. - Core 'LegacyAgentSession' now passes the rich description through 'tool_update' events to ensure dynamic descriptions are updated during execution. - UI 'useAgentStream' hook now extracts these rich values from event metadata to populate the local 'trackedTools' state, allowing 'mapToDisplay' to correctly trigger bespoke rendering components (like the diff viewer).	2026-03-27 15:02:25 -07:00
Michael Bleigh	6975224e45	feat: implement tool display parity for agent stream This commit achieves visual parity for tool execution in the interactive stream when using the experimental 'useAgentProtocol' flag. It removes direct UI dependency on the tool scheduler's internal state. Key changes: - Core 'LegacyAgentSession' now attaches display metadata (displayName, description, etc.) to 'tool_request' AgentEvents. - Core 'LegacyAgentSession' listens to the MessageBus to emit 'tool_update' AgentEvents for live output (e.g., shell commands). - UI 'useAgentStream' now maintains its own 'trackedTools' local state, constructed entirely from incoming 'tool_request', 'tool_update', and 'tool_response' events. - The local 'trackedTools' state is mapped to 'pendingToolGroupItems' using the existing 'mapToDisplay' function for seamless visual parity.	2026-03-27 15:02:18 -07:00
Michael Bleigh	d3a7fc39e3	feat: implement experimental useAgentStream with unified agent protocol This change introduces an experimental 'useAgentProtocol' flag to interactive mode. When enabled, the UI uses the new 'useAgentStream' hook which leverages the core 'LegacyAgentSession' (AgentProtocol) instead of the custom 'useGeminiStream' logic. Key changes: - Added 'useAgentProtocol' experimental setting to CLI and Core config. - Implemented 'useAgentStream' hook with basic interaction and thought support. - Modified 'useToolScheduler' to expose its internal Scheduler instance to ensure implementation parity. - Updated 'AppContainer' to conditionally branch between implementations via ternary operator. - Added comprehensive unit tests for the new hook.	2026-03-27 14:55:11 -07:00
Michael Bleigh	3b317f9198	feat: add experimental useAgentProtocol flag	2026-03-27 14:55:11 -07:00
Sandy Tao	4034c030e7	feat(context): add configurable memoryBoundaryMarkers setting (#24020 )	2026-03-27 21:51:32 +00:00
Gal Zahavi	765fb67011	Enable network access in sandbox configuration (#24055 )	2026-03-27 21:28:01 +00:00
Jacob Richman	97c99f263a	Fix test flakes by globally mocking ink-spinner (#24044 )	2026-03-27 21:19:46 +00:00
gemini-cli-robot	f1a3c35dee	Changelog for v0.36.0-preview.5 (#24046 ) 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>	2026-03-27 21:19:26 +00:00
Jacob Richman	ebe98fdee9	Debug command. (#23851 ) Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>	2026-03-27 21:05:22 +00:00
Abhi	ba71ffa736	fix(core): switch to subshells for shell tool wrapping to fix heredocs and edge cases (#24024 )	2026-03-27 20:34:39 +00:00
joshualitt	320c8aba4c	feat(core): Land `AgentHistoryProvider`. (#23978 )	2026-03-27 19:22:35 +00:00
Sri Pasumarthi	e7dccabf14	fix(acp): prevent crash on empty response in ACP mode (#23952 )	2026-03-27 18:49:13 +00:00
Jacob Richman	a84d4d876e	Increase memory limited for eslint. (#24022 )	2026-03-27 18:40:26 +00:00
Michael Bleigh	6a80311a37	Merge branch 'main' into pr4-cli	2026-03-27 11:08:15 -07:00
Emily Hedlund	29031ea7cf	refactor(core): improve ignore resolution and fix directory-matching bug (#23816 )	2026-03-27 17:12:26 +00:00
Spencer	f3977392e6	revert(core): support custom base URL via env vars (#23976 )	2026-03-27 17:10:14 +00:00
Emily Hedlund	535667baf6	feat(core): add forbiddenPaths to GlobalSandboxOptions and refactor createSandboxManager (#23936 )	2026-03-27 16:57:26 +00:00
Tommaso Sciortino	33cf2da1df	disable extension-reload test (#24018 )	2026-03-27 16:19:15 +00:00
Abhi	104587bae8	feat(core): subagent isolation and cleanup hardening (#23903 )	2026-03-27 03:43:39 +00:00
Yuna Seol	aca8e1af05	fix(core): explicitly set error names to avoid bundling renaming issues (#23913 )	2026-03-27 03:40:49 +00:00
Sam Roberts	6f92642524	ACP integration documents (#22254 ) Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> Co-authored-by: nmcnamara-eng <118702206+nmcnamara-eng@users.noreply.github.com>	2026-03-27 02:10:29 +00:00
gemini-cli-robot	8413dd62ef	Changelog for v0.35.2 (#23960 ) Co-authored-by: gemini-cli-robot <224641728+gemini-cli-robot@users.noreply.github.com>	2026-03-27 02:10:03 +00:00
Sri Pasumarthi	750dec5d8d	fix(acp): Update allow approval policy flow for ACP clients to fix config persistence and compatible with TUI (#23818 )	2026-03-27 01:42:17 +00:00
gemini-cli-robot	335b36893b	Changelog for v0.36.0-preview.4 (#23935 ) Co-authored-by: gemini-cli-robot <224641728+gemini-cli-robot@users.noreply.github.com>	2026-03-27 01:08:39 +00:00
Jenna Inouye	25a20f8e4e	Docs: Update 'docs-writer' skill for relative links (#21463 )	2026-03-27 00:32:20 +00:00
Jacob Richman	b5ba88b001	dep(update) Update Ink version to 6.5.0 (#23843 )	2026-03-26 23:49:51 +00:00
Tommaso Sciortino	8868b34c75	refactor(core): delegate sandbox denial parsing to SandboxManager (#23928 )	2026-03-26 22:10:15 +00:00
Aditya Bijalwan	73dd7328df	feat(core): implement persistent browser session management (#21306 ) Co-authored-by: Gaurav <39389231+gsquared94@users.noreply.github.com> Co-authored-by: cynthialong0-0 <82900738+cynthialong0-0@users.noreply.github.com>	2026-03-26 21:33:37 +00:00
Spencer	d25ce0e143	fix(core): remove shell outputChunks buffer caching to prevent memory bloat and sanitize prompt input (#23751 )	2026-03-26 21:16:07 +00:00
David Pierce	30397816da	feat(sandbox): implement secret visibility lockdown for env files (#23712 ) Co-authored-by: Tommaso Sciortino <sciortino@gmail.com>	2026-03-26 20:35:21 +00:00
Gen Zhang	84f1c19265	feat(cli): enable notifications cross-platform via terminal bell fallback (#21618 ) Co-authored-by: Sandy Tao <sandytao520@icloud.com>	2026-03-26 20:10:49 +00:00
Gal Zahavi	d33170931c	fix(core): allow disabling environment variable redaction (#23927 )	2026-03-26 20:04:44 +00:00
Jenna Inouye	1d230dbfbf	Docs: Update quotas and pricing (#23835 )	2026-03-26 19:29:37 +00:00
Sehoon Shon	c92ae8a359	feat(core): define TrajectoryProvider interface (#23050 )	2026-03-26 19:24:06 +00:00
Keith Schaab	bf03543bf6	fix(a2a-server): A2A server should execute ask policies in interactive mode (#23831 )	2026-03-26 19:10:18 +00:00
matt korwel	1d2fbbf9c3	feat(gcp): add development worker infrastructure (#23814 )	2026-03-26 19:01:37 +00:00
Adib234	9762bf2965	fix(plan): after exiting plan mode switches model to a flash model (#23885 )	2026-03-26 18:45:03 +00:00
ruomeng	c888da5f73	fix(core): replace hardcoded non-interactive ASK_USER denial with explicit policy rules (#23668 )	2026-03-26 18:35:12 +00:00
Dev Randalpura	aa4d9316a9	feat(core): new skill to look for duplicated code while reviewing PRs (#23704 )	2026-03-26 18:32:30 +00:00
Aditya Bijalwan	5755ec2dcf	fix(browser): keep input blocker active across navigations (#22562 ) Co-authored-by: cynthialong0-0 <82900738+cynthialong0-0@users.noreply.github.com>	2026-03-26 16:54:49 +00:00
gemini-cli-robot	a3c1c659fd	Changelog for v0.35.1 (#23840 ) 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>	2026-03-26 16:43:23 +00:00
Adam Weidman	6fb7bcf868	fix(core): preserve first-turn display content	2026-03-26 11:51:15 -04:00
Sehoon Shon	49534209f2	fix(cli): prioritize primary name matches in slash command search (#23850 )	2026-03-26 12:18:57 +00:00
Chris Williams	9e7f52b8f5	Merge examples of use into quickstart documentation (#23319 )	2026-03-26 02:57:23 +00:00
Gal Zahavi	30e0ab102a	feat(sandbox): dynamic Linux sandbox expansion and worktree support (#23692 ) Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>	2026-03-26 01:58:45 +00:00
Alisa	2e03e3aed5	feat(evals): add reliability harvester and 500/503 retry support (#23626 )	2026-03-26 01:48:45 +00:00
Adam Weidman	56656dfbc9	fix(cli): preserve max-turns non-interactive parity	2026-03-23 18:47:11 -04:00
Adam Weidman	93322430a0	fix(cli): align non-interactive loop warning handling	2026-03-23 18:38:56 -04:00
Adam Weidman	25c5d44678	!feat(cli): harden non-interactive agent session handling	2026-03-23 18:38:56 -04:00
Adam Weidman	2dfe237738	!feat(cli): address non-interactive review follow-ups	2026-03-23 18:38:56 -04:00
Adam Weidman	f77e4716fa	!feat(cli): adopt protocol-backed agent session streaming	2026-03-23 18:38:56 -04:00
Adam Weidman	3e0b8aa958	fix(cli): consume the stream returned by send	2026-03-23 18:38:56 -04:00
Adam Weidman	b268e93a1d	refactor(cli): handle max turns from stream end	2026-03-23 18:38:56 -04:00
Adam Weidman	218adcd6c3	feat(cli): migrate nonInteractiveCli to LegacyAgentSession	2026-03-23 18:38:56 -04:00