deps-health-inline

Dependency Health Check (Inline Orchestration)

You ARE the orchestrator. Execute this workflow directly without spawning a separate orchestrator agent.

Workflow Overview

Audit → Validate → Update by Priority → Verify → Repeat if needed

Max iterations: 3 Priorities: critical → high → medium → low

---

Phase 1: Pre-flight

1. Setup directories:

   mkdir -p .tmp/current/{plans,changes,backups}
   

2. Validate environment:

   • Check package.json exists
   • Check type-check and build scripts exist
   • Check lockfile exists (pnpm-lock.yaml, package-lock.json, yarn.lock)

3. Initialize TodoWrite:

   [
     {"content": "Dependency audit", "status": "in_progress", "activeForm": "Auditing dependencies"},
     {"content": "Fix critical dependency issues", "status": "pending", "activeForm": "Fixing critical deps"},
     {"content": "Fix high priority dependency issues", "status": "pending", "activeForm": "Fixing high deps"},
     {"content": "Fix medium priority dependency issues", "status": "pending", "activeForm": "Fixing medium deps"},
     {"content": "Fix low priority dependency issues", "status": "pending", "activeForm": "Fixing low deps"},
     {"content": "Verification audit", "status": "pending", "activeForm": "Verifying updates"}
   ]
   

---

Phase 2: Detection

Invoke dependency-auditor via Task tool:

subagent_type: "dependency-auditor"
description: "Audit all dependencies"
prompt: |
  Audit the entire codebase for dependency issues:
  - Security vulnerabilities (npm audit / pnpm audit)
  - Outdated packages (major/minor/patch)
  - Unused dependencies (via Knip)
  - Deprecated packages
  - License compliance issues
  - Categorize by priority (critical/high/medium/low)

  Generate: dependency-scan-report.md

  Return summary with issue counts per priority.

After dependency-auditor returns:

1. Read dependency-scan-report.md
2. Parse issue counts by priority
3. If zero issues → skip to Final Summary
4. Update TodoWrite: mark audit complete

---

Phase 3: Quality Gate (Detection)

Run inline validation:

pnpm type-check
pnpm build

• If both pass → proceed to updates
• If fail → report to user, exit

---

Phase 4: Update Loop

For each priority (critical → high → medium → low):

1. Check if issues exist for this priority

   • If zero → skip to next priority

2. Update TodoWrite: mark current priority in_progress

3. Invoke dependency-updater via Task tool:

   subagent_type: "dependency-updater"
   description: "Update {priority} dependencies"
   prompt: |
     Read dependency-scan-report.md and fix all {priority} priority issues.
   
     For each issue:
     1. Backup package.json and lockfile
     2. Update ONE dependency at a time
     3. Run type-check and build after each update
     4. If fails, rollback and skip
     5. Log change to .tmp/current/changes/deps-changes.json
   
     Generate/update: dependency-updates-implemented.md
   
     Return: count of updated deps, count of failed updates.
   

4. Quality Gate (inline):

   pnpm type-check
   pnpm build
   

   • If FAIL → report error, suggest rollback, exit
   • If PASS → continue

5. Update TodoWrite: mark priority complete

6. Repeat for next priority

---

Phase 5: Verification

After all priorities updated:

1. Update TodoWrite: mark verification in_progress

2. Invoke dependency-auditor (verification mode):

   subagent_type: "dependency-auditor"
   description: "Verification audit"
   prompt: |
     Re-audit dependencies after updates.
     Compare with previous dependency-scan-report.md.
   
     Report:
     - Issues fixed (count)
     - Issues remaining (count)
     - New issues introduced (count)
   

3. Decision:

   • If issues_remaining == 0 → Final Summary
   • If iteration < 3 AND issues_remaining > 0 → Go to Phase 2
   • If iteration >= 3 → Final Summary with remaining issues

---

Phase 6: Final Summary

Generate summary for user:

## Dependency Health Check Complete

**Iterations**: {count}/3
**Status**: {SUCCESS/PARTIAL}

### Results
- Found: {total} dependency issues
- Fixed: {fixed} ({percentage}%)
- Remaining: {remaining}

### By Priority
- Critical: {fixed}/{total}
- High: {fixed}/{total}
- Medium: {fixed}/{total}
- Low: {fixed}/{total}

### Validation
- Type Check: {status}
- Build: {status}

### Artifacts
- Audit: `dependency-scan-report.md`
- Updates: `dependency-updates-implemented.md`

---

Error Handling

If quality gate fails:

Rollback available: .tmp/current/changes/deps-changes.json

To rollback:
1. Read changes log
2. Restore package.json and lockfile from .tmp/current/backups/
3. Run pnpm install
4. Re-run workflow

If worker fails:

• Report error to user
• Suggest manual intervention
• Exit workflow

---

Key Differences from Old Approach

Old (Orchestrator Agent)	New (Inline Skill)
9+ orchestrator calls	0 orchestrator calls
~1400 lines (cmd + agent)	~150 lines
Context reload each call	Single session context
Plan files for each phase	Direct execution
~10,000+ tokens overhead	~500 tokens

---

Worker Prompts

See references/worker-prompts.md for detailed prompts.