tech-debt-narrator
Tech-debt narrator: finds suppress comments, TODOs, empty catches, workarounds, legacy imports, and type-loosening patterns. Clusters them into logical groups and narrates repayment strategies with ef
SKILL.mdskills/understanding/tech-debt-narrator/SKILL.md
GitHubWhat this is for
eslint-disable, TODO from 2019, empty catches, as any - listed as a flat checklist they are noise. Clustered by logical group they become actionable: "auth-middleware has 3 related debt items, fixing them together costs S not M." This skill finds, clusters, and narrates a repayment strategy.
PROTECTION RULE - never ~/.claude/
Read-only skill. Guard required if write mode added later.
## What You Must Do When Invoked
During analysis, assign a confidence level to each finding: proven (confirmed by evidence), likely (strong signal, needs review), or suspected (weak signal).
Step 1
- If
-helpis passed, print the## Usageblock below and stop.
Step 2
- Confirm
-ProjectDiris provided and the path exists.
Step 3
- Run:
scripts/debt-scan.ps1 -ProjectDir "<path>"
Step 4
- LLM reads the JSON output. For each finding in
debts[]:
Step 5
- Suppress/todo/empty-catch/workaround/legacy/type-loosen: is this genuine debt or legitimate usage? (
@ts-ignorenext to a known-unsound interop vs one hiding a type error).
Step 6
- Cluster debts into logical groups (e.g. all items in
auth-middlewarethat touch the same function). Per cluster:
Step 7
- Narrative description: what is the debt, why was it introduced, what breaks if not fixed.
Step 8
- Repayment strategy: refactor / rewrite / retire.
Step 9
- Estimated effort: S / M / L.
Step 10
- Write
tech-debt-report.mdto the working directory.
Usage
/tech-debt # interactive
/tech-debt <dir> # scan project
/tech-debt -help # show usage
Returns JSON with debts[] (file, line, type, text, ageDays, severity), counts{} plus cluster hints.