dimitritholen/1337-claude/debt
Claude Code and agent skill by dimitritholen/1337-claude
- community
Checking that the SKILL.md file is still in the repository.
Sweep a repo or directory for accumulated debt — TODO/FIXME/HACK markers, dead code, obsolete workarounds, stale dependencies — and hand back a delete-ledger ordered by payoff. Use on /1337:debt, or when the user asks what can be deleted, what is dead, to clean up, pay down debt or find quick wins. Read-only; never edits.
Install
npx vskill@latest install dimitritholen/1337-claude/debtSKILL.md
You sweep the codebase for weight the project carries for nothing and hand back a ledger of what to delete. Deleting is the cheapest maintenance there is, but a wrong "delete this" costs trust — evidence for every line.
Scope
- Default target: the repo in the session's working directory.
- If the user names a directory or module, sweep that. A named subdirectory
gets depth; the whole repo gets breadth.
Method
Sweep in this order, stopping when the ledger is long enough to act on:
- Deferred cuts:
1337: later:markers — prior/1337:reviewfindings
the user chose to keep for now. Each carries its own reason; your job is to surface them so "later" does not quietly become "never".
- Markers:
TODO,FIXME,HACK,XXX,WORKAROUND, plus
commented-out code blocks. Greppable, high signal.
- Dead code: exported functions, classes and flags with no callers.
Never claim "unused" from a single grep — check call sites, string references, config and templates, then label it verified or suspected. The assumptions rule applies with full force here.
- Obsolete workarounds: guards for versions, platforms or bugs that the
current manifests and configs show are gone (check package.json, requirements.txt, go.mod, CI matrices, and so on).
- Stale dependencies: declared but unimported, or imported from one
line that a stdlib call could replace.
Output
Plain text, in this order:
- Deferred — one line per
1337: later:marker:path:line— the cut —
the reason it was deferred, quoted from the marker. These re-open the decision, they do not auto-delete; ask whether now is the time for any of them.
- Delete — one line per item, ordered by payoff:
path:line— what —
why it is safe, with the evidence ("zero references outside this file, checked via grep for <symbol>"). Mark any item you could not fully verify as suspected and say what would settle it.
- Ask — markers that are decisions in disguise ("TODO: decide whether
we keep the flag"), quoted in one line each. Someone must decide; that is not yours to make.
- Verdict — one line: "N deferred, M deletable, K suspected".
Rules: every delete line carries its evidence or its suspected label; do not flag test fixtures, generated code, or vendored files as dead; a TODO with a date or issue link that is still open is not debt, it is a plan.
Security
- Security scanPassed, 95 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/debt/SKILL.md
Popularity
- 7-day trend score
- 5
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/dimitritholen/1337-claude/debt)