Bug Health Check (Inline Orchestration)

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

Workflow Overview

Beads Init → Detection → History Check (HIGH+) → Create Issues → Fix by Priority → Close Issues → Verify → Beads Complete

Max iterations: 3 Priorities: critical → high → medium → low Beads integration: Automatic issue tracking History enrichment: For CRITICAL and HIGH bugs only

Phase 1: Pre-flight & Beads Init

1. Setup directories:

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

2. Validate environment:

   • Check package.json exists
   • Check type-check and build scripts exist

3. Create Beads wisp:

   bd mol wisp healthcheck

   IMPORTANT: Save the wisp ID (e.g., mc2-xxx) for later use.

4. Initialize TodoWrite:

   [
     { "content": "Bug detection", "status": "in_progress", "activeForm": "Detecting bugs" },
     { "content": "Create Beads issues", "status": "pending", "activeForm": "Creating issues" },
     { "content": "Fix critical bugs", "status": "pending", "activeForm": "Fixing critical bugs" },
     { "content": "Fix high priority bugs", "status": "pending", "activeForm": "Fixing high bugs" },
     {
       "content": "Fix medium priority bugs",
       "status": "pending",
       "activeForm": "Fixing medium bugs"
     },
     { "content": "Fix low priority bugs", "status": "pending", "activeForm": "Fixing low bugs" },
     { "content": "Verification scan", "status": "pending", "activeForm": "Verifying fixes" },
     { "content": "Complete Beads wisp", "status": "pending", "activeForm": "Completing wisp" }
   ]

Phase 2: Detection

Invoke bug-hunter via Task tool:

subagent_type: "bug-hunter"
description: "Detect all bugs"
prompt: |
  Scan the entire codebase for bugs:
  - Run type-check and build
  - Check for security vulnerabilities
  - Find dead code and debug statements
  - Categorize by priority (critical/high/medium/low)

  Generate: bug-hunting-report.md

  Return summary with bug counts per priority.

After bug-hunter returns:

1. Read bug-hunting-report.md
2. Parse bug counts by priority
3. If zero bugs → skip to Phase 7 (Final Summary)
4. Update TodoWrite: mark detection complete

Phase 2.5: History Enrichment (CRITICAL/HIGH only)

Purpose: Find previously fixed similar bugs to detect regressions and provide historical context.

When to run: Only for CRITICAL and HIGH priority bugs (skip for MEDIUM/LOW).

Steps

1. Extract keywords from each CRITICAL/HIGH bug title/description

2. Search Beads history:

   # For each bug, search by relevant keywords
   bd search "{keywords}" --status closed --limit 5
   
   # Also search by category
   bd search "security" --status closed --limit 5      # for security bugs
   bd search "dependency" --status closed --limit 5    # for dependency bugs
   bd search "{file_path}" --status closed --limit 3   # for file-specific bugs

3. Evaluate results:

   • Match found: Similar closed issue exists
     • Could be regression (same bug returned)
     • Could be related pattern (similar problem elsewhere)
   • No match: New type of bug

4. Enrich bug data:

   For each bug with history matches:
   - Add to bug metadata: related_issues: [mc2-xxx, mc2-yyy]
   - Note if potential regression: is_regression: true/false

5. Store enrichment for Phase 3 and Phase 5:

   // .tmp/current/history-enrichment.json
   {
     "bug_1": {
       "related_closed": ["mc2-abc", "mc2-def"],
       "is_potential_regression": false,
       "context": "Similar security fix in mc2-abc"
     }
   }

Output

• If related issues found: Include in bug description when creating Beads issue
• If potential regression: Flag for special attention in bug-fixer prompt
• If no history: Proceed normally (new bug type)

Example

# Bug: "Vulnerable tar package"
bd search "tar" --status closed --limit 5
bd search "vulnerability" --status closed --limit 5
bd search "pnpm override" --status closed --limit 5

# Results: No matches → new bug, no enrichment needed
# Results: mc2-xyz found → add to related_issues

Note: History enrichment is informational, not blocking. Missing history doesn't prevent fixing.

Phase 3: Create Beads Issues

For each bug found, create a Beads issue.

For CRITICAL/HIGH bugs (with history enrichment):

# If related_issues found in Phase 2.5:
bd create "BUG: {bug_title}" -t bug -p {1|2} \
  -d "{description}

## History Context
Related closed issues: {related_issues}
Potential regression: {yes/no}
Previous fix context: {context from enrichment}" \
  --deps discovered-from:{wisp_id}

# If no history found:
bd create "BUG: {bug_title}" -t bug -p {1|2} -d "{description}" \
  --deps discovered-from:{wisp_id}

For MEDIUM/LOW bugs (no history check):

# Medium bugs (P3)
bd create "BUG: {bug_title}" -t bug -p 3 -d "{description}" \
  --deps discovered-from:{wisp_id}

# Low bugs (P4)
bd create "BUG: {bug_title}" -t bug -p 4 -d "{description}" \
  --deps discovered-from:{wisp_id}

Track issue IDs in a mapping:

bug_1 → mc2-aaa (related: mc2-xyz)
bug_2 → mc2-bbb
...

Update TodoWrite: mark "Create Beads issues" complete.

Phase 4: Quality Gate (Pre-fix)

Run inline validation:

pnpm type-check
pnpm build

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

Phase 5: Fixing Loop

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

1. Check if bugs exist for this priority

   • If zero → skip to next priority

2. Update TodoWrite: mark current priority in_progress

3. Claim issues in Beads:

   bd update {issue_id} --status in_progress

4. Invoke bug-fixer via Task tool:

   subagent_type: "bug-fixer"
   description: "Fix {priority} bugs"
   prompt: |
     Read bug-hunting-report.md and fix all {priority} priority bugs.
   
     ## History Context (for CRITICAL/HIGH only)
     Check .tmp/current/history-enrichment.json for related closed issues.
     If a bug has related_closed issues:
     - Review the previous fix approach (bd show {related_id})
     - Consider if this is a regression
     - Apply learnings from previous fix
   
     For each bug:
     1. Backup file before editing
     2. If history exists, review previous fix first
     3. Implement fix
     4. Log change to .tmp/current/changes/bug-changes.json
   
     Generate/update: bug-fixes-implemented.md
   
     Return: count of fixed bugs, count of failed fixes, list of fixed bug IDs.

5. Quality Gate (inline):

   pnpm type-check
   pnpm build

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

6. Close fixed issues in Beads:

   bd close {issue_id_1} {issue_id_2} ... --reason "Fixed in health check"

7. Update TodoWrite: mark priority complete

8. Repeat for next priority

Phase 6: Verification

After all priorities fixed:

1. Update TodoWrite: mark verification in_progress

2. Invoke bug-hunter (verification mode):

   subagent_type: "bug-hunter"
   description: "Verification scan"
   prompt: |
     Re-scan codebase after fixes.
     Compare with previous bug-hunting-report.md.
   
     Report:
     - Bugs fixed (count)
     - Bugs remaining (count)
     - New bugs introduced (count)

3. Decision:

   • If bugs_remaining == 0 → Phase 7
   • If iteration < 3 AND bugs_remaining > 0 → Go to Phase 2
   • If iteration >= 3 → Phase 7 with remaining bugs

Phase 7: Final Summary & Beads Complete

1. Complete Beads wisp:

   # If all bugs fixed
   bd mol squash {wisp_id}
   
   # If no bugs found (nothing to do)
   bd mol burn {wisp_id}

2. Create issues for remaining bugs (if any):

   bd create "REMAINING: {bug_title}" -t bug -p {priority} \
     -d "Not fixed in health check. See bug-hunting-report.md"

3. Generate summary for user:

## Bug Health Check Complete

**Wisp ID**: {wisp_id}
**Iterations**: {count}/3
**Status**: {SUCCESS/PARTIAL}

### Results

- Found: {total} bugs
- Fixed: {fixed} ({percentage}%)
- Remaining: {remaining}

### By Priority

- Critical: {fixed}/{total}
- High: {fixed}/{total}
- Medium: {fixed}/{total}
- Low: {fixed}/{total}

### Beads Issues

- Created: {count}
- Closed: {count}
- Remaining: {count} (issues created for follow-up)

### Validation

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

### Artifacts

- Detection: `bug-hunting-report.md`
- Fixes: `bug-fixes-implemented.md`

4. Update TodoWrite: mark wisp complete

5. SESSION CLOSE PROTOCOL:

   git status
   git add .
   bd sync
   git commit -m "fix: health check - {fixed} bugs fixed ({wisp_id})"
   bd sync
   git push

Error Handling

If quality gate fails:

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

To rollback:
1. Read changes log
2. Restore files from .tmp/current/backups/
3. Re-run workflow

If worker fails:

• Report error to user
• Keep Beads wisp open for manual completion
• Suggest manual intervention
• Exit workflow

If Beads command fails:

• Log error but continue workflow
• Beads tracking is enhancement, not blocker

Quick Reference

Worker Prompts

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