feat(core): integrate skill-creator into skill extraction agent (#25421)

evals/skill_extraction.eval.ts

@@ -272,8 +272,12 @@ describe('Skill Extraction', () => {
       expect(combinedSkills).toContain('npm run predocs:settings');
       expect(combinedSkills).toContain('npm run schema:settings');
       expect(combinedSkills).toContain('npm run docs:settings');
-      expect(combinedSkills).toMatch(/When to Use/i);
       expect(combinedSkills).toMatch(/Verification/i);
+
+      // Verify the extraction agent activated skill-creator for design guidance.
+      expect(config.getSkillManager().isSkillActive('skill-creator')).toBe(
+        true,
+      );
     },
   });
 
@@ -335,7 +339,11 @@ describe('Skill Extraction', () => {
       expect(combinedSkills).toContain('npm run db:migrate');
       expect(combinedSkills).toContain('npm run db:validate');
       expect(combinedSkills).toMatch(/rollback/i);
-      expect(combinedSkills).toMatch(/When to Use/i);
+
+      // Verify the extraction agent activated skill-creator for design guidance.
+      expect(config.getSkillManager().isSkillActive('skill-creator')).toBe(
+        true,
+      );
     },
   });
 });

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

@@ -7,11 +7,13 @@
 import { z } from 'zod';
 import type { LocalAgentDefinition } from './types.js';
 import {
+  ACTIVATE_SKILL_TOOL_NAME,
   EDIT_TOOL_NAME,
   GLOB_TOOL_NAME,
   GREP_TOOL_NAME,
   LS_TOOL_NAME,
   READ_FILE_TOOL_NAME,
+  SHELL_TOOL_NAME,
   WRITE_FILE_TOOL_NAME,
 } from '../tools/tool-names.js';
 import { PREVIEW_GEMINI_FLASH_MODEL } from '../config/models.js';
@@ -152,45 +154,6 @@ function buildSystemPrompt(skillsDir: string): string {
     '- One-off artifact names: bug IDs, branch names, timestamps, exact incident strings',
     '',
     '============================================================',
-    'SKILL FORMAT',
-    '============================================================',
-    '',
-    'Each skill is a directory containing a SKILL.md file with YAML frontmatter',
-    'and optional supporting scripts.',
-    '',
-    'Directory structure:',
-    `  ${skillsDir}/<skill-name>/`,
-    '    SKILL.md            # Required entrypoint',
-    '    scripts/<tool>.*    # Optional helper scripts (Python stdlib-only or shell)',
-    '',
-    'SKILL.md structure:',
-    '',
-    '  ---',
-    '  name: <skill-name>',
-    '  description: <1-2 lines; include concrete triggers in user-like language>',
-    '  ---',
-    '',
-    '  ## When to Use',
-    '  <Clear trigger conditions and non-goals>',
-    '',
-    '  ## Procedure',
-    '  <Numbered steps with specific commands, paths, code patterns>',
-    '',
-    '  ## Pitfalls and Fixes',
-    '  <symptom -> likely cause -> fix; only include observed failures>',
-    '',
-    '  ## Verification',
-    '  <Concrete success checks>',
-    '',
-    'Supporting scripts (optional but recommended when applicable):',
-    '- Put helper scripts in scripts/ and reference them from SKILL.md',
-    '- Prefer Python (stdlib only) or small shell scripts',
-    '- Make scripts safe: no destructive actions, no secrets, deterministic output',
-    '- Include a usage example in SKILL.md',
-    '',
-    'Naming: kebab-case (e.g., fix-lint-errors, run-migrations).',
-    '',
-    '============================================================',
     'UPDATING EXISTING SKILLS (PATCHES)',
     '============================================================',
     '',
@@ -247,20 +210,28 @@ function buildSystemPrompt(skillsDir: string): string {
     '',
     `1. Use list_directory on ${skillsDir} to see existing skills.`,
     '2. If skills exist, read their SKILL.md files to understand what is already captured.',
-    '3. Scan the session index provided in the query. Look for [NEW] sessions whose summaries',
+    '3. Use activate_skill to load the "skill-creator" skill. Follow its design guidance',
+    '   (conciseness, progressive disclosure, frontmatter format, bundled resources) when',
+    '   writing SKILL.md files. You may also use its init_skill.cjs script to scaffold new',
+    '   skill directories and package_skill.cjs to validate finished skills.',
+    '   IMPORTANT: You are a background agent with no user interaction. Skip any interactive',
+    '   steps in the skill-creator guide (asking clarifying questions, requesting user feedback,',
+    '   installation prompts, iteration loops). Use only its format and quality guidance.',
+    '4. Scan the session index provided in the query. Look for [NEW] sessions whose summaries',
     '   hint at workflows that ALSO appear in other sessions (either [NEW] or [old]) or at a',
     '   stable recurring repo workflow. Remember: summary similarity alone is NOT enough.',
-    '4. Apply the minimum signal gate. If recurrence or durability is not visible, report that',
+    '5. Apply the minimum signal gate. If recurrence or durability is not visible, report that',
     '   no skill should be created and finish.',
-    '5. For promising patterns, use read_file on the session file paths to inspect the full',
+    '6. For promising patterns, use read_file on the session file paths to inspect the full',
     '   conversation. Confirm the workflow was actually repeated and validated. Read at least',
     '   two sessions unless the candidate is clearly a stable recurring repo lifecycle workflow.',
-    '6. For each candidate, verify it meets ALL criteria. Before writing, make sure you can',
+    '7. For each candidate, verify it meets ALL criteria. Before writing, make sure you can',
     '   state: future trigger, evidence sessions, recurrence signal, validation signal, and',
     '   why it is not generic.',
-    '7. Write new SKILL.md files or update existing ones in your directory using write_file.',
+    '8. Write new SKILL.md files or update existing ones in your directory.',
+    '   Use run_shell_command to run init_skill.cjs for scaffolding and package_skill.cjs for validation.',
     '   For skills that live OUTSIDE your directory, write a .patch file instead (see UPDATING EXISTING SKILLS).',
-    '8. Write COMPLETE files — never partially update a SKILL.md.',
+    '9. Write COMPLETE files — never partially update a SKILL.md.',
     '',
     'IMPORTANT: Do NOT read every session. Only read sessions whose summaries suggest a',
     'repeated pattern or a stable recurring repo workflow worth investigating. Most runs',
@@ -274,8 +245,9 @@ function buildSystemPrompt(skillsDir: string): string {
  * writes reusable SKILL.md files to the project memory directory.
  *
  * This agent is designed to run in the background on session startup.
- * It has restricted tool access (file tools only, no shell or user interaction)
+ * It has restricted tool access (file tools, shell, and skill activation — no
- * and is prompted to only operate within the skills memory directory.
+ * user interaction) and is prompted to only operate within the skills memory
+ * directory.
  */
 export const SkillExtractionAgent = (
   skillsDir: string,
@@ -309,12 +281,14 @@ export const SkillExtractionAgent = (
   },
   toolConfig: {
     tools: [
+      ACTIVATE_SKILL_TOOL_NAME,
       READ_FILE_TOOL_NAME,
       WRITE_FILE_TOOL_NAME,
       EDIT_TOOL_NAME,
       LS_TOOL_NAME,
       GLOB_TOOL_NAME,
       GREP_TOOL_NAME,
+      SHELL_TOOL_NAME,
     ],
   },
   get promptConfig() {