| name | decide-tech-debt |
| description | Decide whether to pay down a specific tech-debt by sizing its carrying cost and payback against the roadmap — not all-or-nothing. Reach for this before a 'we must stop and refactor' or 'we never have time' reflex. |
Skill: Decide tech-debt
Tech-debt is a business decision with a carrying cost, not a moral failing (§3 #7).
Step 1 — Measure the pain, don't feel it
Lead-time drift, change-fail, rework, and hotspots — separate felt pain from measured pain (§3 #4 #7). "The codebase is bad" is not a finding.
Step 2 — Is it on a hotspot?
High churn × high complexity is where paydown has leverage. Stable code, even ugly, is low priority (§3 #7).
Step 3 — Size carrying cost + payback
Run engineering_management_calc.py tech-debt to compute the carrying cost per period and the payback periods of paying it down (§3 #7).
Step 4 — Trade against the roadmap explicitly
Pay down when payback is short relative to the code's remaining life; prefer incremental (strangler-fig) over a rewrite (the highest-risk, longest-payback option). Reserve a standing capacity slice rather than a one-off crusade.
Output
A tech-debt decision memo: measured pain, hotspot status, sized carrying cost + payback, and a recommendation with a capacity slice. Use ../../templates/tech-debt-decision-memo.md. Traverse Tree 3. The architecture itself routes to ravenclaude-core/architect.
