Tech Debt Triage Skill

Score, prioritize, and plan technical debt remediation.

Trigger Conditions

• Sprint planning session
• Quarterly tech debt review
• User invokes with "triage tech debt" or "debt assessment"

Input Contract

• Required: Codebase metrics (churn, complexity, defect density)
• Optional: Developer pain surveys, postmortem findings, code review history

Output Contract

• Scored and prioritized debt backlog
• Quick wins identified (low effort, high impact)
• Remediation roadmap with effort estimates
• Contagion analysis (debt spreading to new code)

Tool Permissions

• Read: Code metrics, git history, defect tracker, postmortems
• Write: Tech debt reports, backlog items
• Search: Code hotspots and churn patterns

Execution Steps

1. Collect code metrics: churn, complexity, defect density
2. Identify hotspots (high churn + high complexity)
3. Score each debt item: impact (blast radius), effort (team-days), contagion
4. Identify quick wins (effort <1d, impact >medium)
5. Check for recurring patterns from postmortems
6. Create prioritized backlog
7. Estimate remediation effort for top items

Success Criteria

• All debt items scored on 3 dimensions
• Quick wins identified and flagged
• Top 10 items have effort estimates
• Contagion patterns identified

Escalation Rules

• Escalate if debt item blocks feature delivery
• Escalate if contagion is spreading to >3 modules
• Escalate if debt remediation requires >1 sprint of dedicated work

Example Invocations

Input: "Triage tech debt for the order processing module"

Output: 14 debt items identified. Top 3: 1) OrderProcessor god class (impact: 9, effort: 3d, contagion: HIGH — 6 modules depend on it), 2) Missing retry logic in payment integration (impact: 8, effort: 1d, contagion: MEDIUM), 3) Hardcoded config values (impact: 5, effort: 0.5d, contagion: LOW — quick win). Remediation roadmap: sprint N (quick wins), sprint N+1 (retry logic), sprint N+2-N+3 (OrderProcessor decomposition).