feat(plan): enable built-in research subagents in plan mode

docs/cli/plan-mode.md

@@ -25,7 +25,7 @@ implementation. It allows you to:
   - [Customizing Planning with Skills](#customizing-planning-with-skills)
   - [Customizing Policies](#customizing-policies)
     - [Example: Allow git commands in Plan Mode](#example-allow-git-commands-in-plan-mode)
-    - [Example: Enable research subagents in Plan Mode](#example-enable-research-subagents-in-plan-mode)
+    - [Example: Enable custom subagents in Plan Mode](#example-enable-custom-subagents-in-plan-mode)
   - [Custom Plan Directory and Policies](#custom-plan-directory-and-policies)
 - [Automatic Model Routing](#automatic-model-routing)
 - [Cleanup](#cleanup)
@@ -134,6 +134,7 @@ These are the only allowed tools:
 
 - **FileSystem (Read):** [`read_file`], [`list_directory`], [`glob`]
 - **Search:** [`grep_search`], [`google_web_search`]
+- **Research Subagents:** [`codebase_investigator`], [`cli_help`]
 - **Interaction:** [`ask_user`]
 - **MCP Tools (Read):** Read-only [MCP tools] (e.g., `github_read_issue`,
   `postgres_read_schema`) are allowed.
@@ -204,16 +205,17 @@ priority = 100
 modes = ["plan"]
 ```
 
-#### Example: Enable research subagents in Plan Mode
+#### Example: Enable custom subagents in Plan Mode
 
-You can enable experimental research [subagents] like `codebase_investigator` to
-help gather architecture details during the planning phase.
+Built-in research [subagents] like [`codebase_investigator`] and [`cli_help`]
+are enabled by default in Plan Mode. You can enable additional [custom
+subagents] by adding a rule to your policy.
 
 `~/.gemini/policies/research-subagents.toml`
 
 ```toml
 [[rule]]
-toolName = "codebase_investigator"
+toolName = "my_custom_subagent"
 decision = "allow"
 priority = 100
 modes = ["plan"]
@@ -319,7 +321,10 @@ those files are not automatically deleted and must be managed manually.
 [MCP tools]: /docs/tools/mcp-server.md
 [`save_memory`]: /docs/tools/memory.md
 [`activate_skill`]: /docs/cli/skills.md
+[`codebase_investigator`]: /docs/core/subagents.md#codebase_investigator
+[`cli_help`]: /docs/core/subagents.md#cli_help
 [subagents]: /docs/core/subagents.md
+[custom subagents]: /docs/core/subagents.md#creating-custom-subagents
 [policy engine]: /docs/reference/policy-engine.md
 [`enter_plan_mode`]: /docs/tools/planning.md#1-enter_plan_mode-enterplanmode
 [`exit_plan_mode`]: /docs/tools/planning.md#2-exit_plan_mode-exitplanmode

packages/core/src/policy/policies/plan.toml

@@ -72,7 +72,16 @@ priority = 70
 modes = ["plan"]
 
 [[rule]]
-toolName = ["glob", "grep_search", "list_directory", "read_file", "google_web_search", "activate_skill"]
+toolName = [
+  "glob",
+  "grep_search",
+  "list_directory",
+  "read_file",
+  "google_web_search",
+  "activate_skill",
+  "codebase_investigator",
+  "cli_help"
+]
 decision = "allow"
 priority = 70
 modes = ["plan"]

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

@@ -1593,7 +1593,7 @@ describe('PolicyEngine', () => {
           modes: [ApprovalMode.PLAN],
         },
         {
-          toolName: 'codebase_investigator',
+          toolName: 'unknown_subagent',
           decision: PolicyDecision.ALLOW,
           priority: PRIORITY_SUBAGENT_TOOL,
         },
@@ -1605,7 +1605,7 @@ describe('PolicyEngine', () => {
       });
 
       const fixedResult = await fixedEngine.check(
-        { name: 'codebase_investigator' },
+        { name: 'unknown_subagent' },
         undefined,
       );
 

packages/core/src/policy/toml-loader.test.ts

@@ -909,7 +909,7 @@ priority = 100
       }
     });
 
-    it('should override default subagent rules when in Plan Mode', async () => {
+    it('should override default subagent rules when in Plan Mode for unknown subagents', async () => {
       const planTomlPath = path.resolve(__dirname, 'policies', 'plan.toml');
       const fileContent = await fs.readFile(planTomlPath, 'utf-8');
       const tempPolicyDir = await fs.mkdtemp(
@@ -931,9 +931,9 @@ priority = 100
           approvalMode: ApprovalMode.PLAN,
         });
 
-        // 3. Simulate a Subagent being registered (Dynamic Rule)
+        // 3. Simulate an unknown Subagent being registered (Dynamic Rule)
         engine.addRule({
-          toolName: 'codebase_investigator',
+          toolName: 'unknown_subagent',
           decision: PolicyDecision.ALLOW,
           priority: PRIORITY_SUBAGENT_TOOL,
           source: 'AgentRegistry (Dynamic)',
@@ -942,13 +942,13 @@ priority = 100
         // 4. Verify Behavior:
         // The Plan Mode "Catch-All Deny" (from plan.toml) should override the Subagent Allow
         const checkResult = await engine.check(
-          { name: 'codebase_investigator' },
+          { name: 'unknown_subagent' },
           undefined,
         );
 
         expect(
           checkResult.decision,
-          'Subagent should be DENIED in Plan Mode',
+          'Unknown subagent should be DENIED in Plan Mode',
         ).toBe(PolicyDecision.DENY);
 
         // 5. Verify Explicit Allows still work
@@ -958,6 +958,25 @@ priority = 100
           readResult.decision,
           'Explicitly allowed tools (read_file) should be ALLOWED in Plan Mode',
         ).toBe(PolicyDecision.ALLOW);
+
+        // 6. Verify Built-in Research Subagents are ALLOWED
+        const codebaseResult = await engine.check(
+          { name: 'codebase_investigator' },
+          undefined,
+        );
+        expect(
+          codebaseResult.decision,
+          'codebase_investigator should be ALLOWED in Plan Mode',
+        ).toBe(PolicyDecision.ALLOW);
+
+        const cliHelpResult = await engine.check(
+          { name: 'cli_help' },
+          undefined,
+        );
+        expect(
+          cliHelpResult.decision,
+          'cli_help should be ALLOWED in Plan Mode',
+        ).toBe(PolicyDecision.ALLOW);
       } finally {
         await fs.rm(tempPolicyDir, { recursive: true, force: true });
       }