Regular debt review with stale item cleanup

Severity: Recommended

A debt list without regular review becomes a graveyard of good intentions. Reviews keep the list relevant, prioritized, and actionable.

Check automatically:

# Look for review meeting notes or cadence docs
grep -riE "debt review|quarterly review|backlog grooming|tech debt.*meeting" docs/ --include="*.md" 2>/dev/null

# Check for process documentation
find . -maxdepth 4 -type f \( -name "*process*" -o -name "*cadence*" -o -name "*rituals*" \) -name "*.md" 2>/dev/null | grep -v node_modules

# Find stale debt items (older than 1 year) if using GitHub
gh issue list --label "tech-debt" --state open --json title,createdAt --limit 50 2>/dev/null | jq '[.[] | select(.createdAt < "2025-02-01")] | length'

Ask user:

• "Is there a scheduled cadence for reviewing tech debt?" (quarterly, monthly, sprint planning)
• "When was the last debt review?"
• "Who participates?" (just leads, or whole team)
• "What happens in the review?" (prioritize, close stale items, add new ones)
• "Are there debt items older than 1 year still open?"

Review should include:

• Prioritization of open items
• Closing stale or irrelevant items
• Adding newly discovered debt
• Validating effort estimates
• Deciding what to tackle next

Pass criteria:

• Defined cadence exists (quarterly minimum)
• Review actually happens (not just documented)
• Last review within expected window
• Stale items addressed (closed or deliberately kept)
• No debt items older than 12 months without explicit decision

Fail criteria:

• No scheduled review ("we look at it when we have time")
• Cadence documented but not followed
• Last review was 6+ months ago
• Debt items from 2+ years ago still open and ignored
• Backlog only grows, never shrinks

Cross-reference with:

• DEBT-001 (review validates the list is alive)
• DEBT-004 (review is where allocation gets decided)

Evidence to capture:

• Review cadence (documented vs actual)
• Date of last review
• Count of stale items (>12 months old)
• Actions taken in recent review

---