SkillsAdd

laststance/skills

codebase-litter-audit

Find unfinished code litter

First seen May 17, 2026

Installation

$ npx skills add laststance/skills --skill codebase-litter-audit
View on skills.sh ↗ GitHub repository ↗

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

extension-to-functions-codebase Similar name
firebase/agent-skills

Skill for converting an installed Firebase Extension (or extension source) into a standalone Cl…

30K installs
generate-custom-instructions-from-codebase Similar name
github/awesome-copilot

Migration and code evolution instructions generator for GitHub Copilot. Analyzes differences be…

9K installs
acquire-codebase-knowledge Similar name
github/awesome-copilot

Use this skill when the user explicitly asks to map, document, or onboard into an existing code…

1.8K installs
planetscale-codebase-sqlcommenter-instrumentation Similar name
planetscale/skills

Inspect an application repository connected to PlanetScale and recommend SQLCommenter-compatibl…

13 installs
improve-codebase-architecture Similar name
mattpocock/skills 891.6K installs
codebase-design Similar name
mattpocock/skills

Shared vocabulary for designing deep modules. Use when the user wants to design or improve a mo…

575K installs

Also in this package

Other skills from laststance/skills · top by installs.

npx skills add laststance/skills
coderabbit-resolver 51 product-inspiration 49 qa-team 47 electron-release 46 git 45 code-trace 45 load 43 save 42 explain 41 task 40

Browse all from laststance/skills

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Declared
Cursor Declared
Codex Declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Repository health

Stars 1
License LICENSE
Default branch main
Open issues 3
Status Active

Skill metadata

Parsed from SKILL.md frontmatter.

Declared agents claude-code cursor codex

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 4,766 B
  • docs SUMMARY.md 56 B

History

  1. First seen on skills.sh
  2. First recorded snapshot · 10 installs

SKILL.md

Codebase Litter Audit

Codex Compatibility

When running this skill in Codex, translate Claude Code-only primitives before acting: AskUserQuestion -> chat/requestuserinput, TodoWrite -> updateplan, Task/TaskCreate/TeamCreate/SendMessage -> spawnagent/sendinput/waitagent when available and allowed, and EnterPlanMode/ExitPlanMode -> a concise chat plan plus explicit approval. Resolve Read/Write/Edit/Bash/WebSearch/WebFetch to Codex file/shell/web tools, and map ~/.claude/... paths to ~/.agents/... or ~/.codex/... unless the task explicitly targets Claude Code.

Cursor Compatibility

When running this skill in Cursor Agent, translate Claude Code-only primitives before acting: AskUserQuestion -> AskQuestion; TodoWrite -> Cursor TodoWrite or an equivalent checklist; Task/TaskCreate/TeamCreate/SendMessage/multi-agent flows -> Cursor Task (subagents), parallel Tasks, or runinbackground when allowed (TeamCreate/SendMessage may have no exact match); EnterPlanMode/ExitPlanMode -> Plan mode (SwitchMode / CreatePlan) plus explicit approval. Resolve Read/Write/Edit/StrReplace/Bash/web/search/MCP via Cursor Composer or Agent equivalents. MCP names written as mcpservertool typically map to callmcptool with configured server identifiers. Map ~/.claude/... to ~/.cursor/skills/, .cursor/skills/, and .cursor/rules/ unless the task explicitly targets Claude Code.

Overview

Find incomplete or misleading work that is still reachable, visible, or confusing even when dead-code scanners report nothing. Treat the scanner as a lead generator, then verify each finding by reading neighboring code, tests, docs, and issue tracker context.

Workflow

  1. Establish scope.

- Run git status --short and note whether the worktree is dirty. - Identify the repo root and primary tracker (gh issue list when GitHub is configured). - Do not edit files unless the user explicitly asks for cleanup.

  1. Collect candidate signals.

- Prefer the bundled scanner:

python3 ~/.codex/skills/codebase-litter-audit/scripts/scan_litter.py . --markdown

- If the repo is huge, pass focused paths after the repo root: --include src --include e2e. - Supplement with rg for domain words the scanner cannot infer, such as feature names, issue IDs, or product terminology.

  1. Verify candidates manually.

- Read the candidate line and nearby code. - Trace one call site, UI surface, test, or doc reference before calling it real. - Search for later migrations, docs, PRs, or issues that may supersede an old TODO. - Classify stale comments separately from active defects.

  1. Report with evidence.

- Lead with high-confidence, user-visible, or security-relevant litter. - Anchor every claim to a clickable file path and line. - Include severity as High, Medium, or Low. - State the smallest cleanup path: remove UI, implement behavior, unskip/replace test, update docs, or close as stale. - Mention likely duplicate/open issues if found.

What Counts

Use references/signal-catalog.md when deciding whether a candidate is worth reporting.

Strong signals:

  • Visible commands, routes, buttons, menus, settings, or CLI options wired to stubs or no-op handlers.
  • TODO, FIXME, not implemented, stub, temporary, or WIP attached to reachable product code.
  • Tests skipped for real product behavior, especially persistence, auth, security, data loss, or critical flows.
  • Comments or docs that contradict later code, migrations, or configuration.
  • Placeholder assets or copy that can ship to users.
  • Suppressions such as eslint-disable, ts-ignore, @ts-expect-error, or broad ignores with no issue link or expiry.

Weak signals:

  • Example text in stories/tests.
  • Placeholder attributes in normal forms.
  • Historical migration comments already superseded by a later migration.
  • Generated files, lockfiles, vendored code, or .git hook samples.

Output Template

## Codebase Litter Audit

### Findings

- **Medium: <short title>**
  Evidence: [file.ts](/abs/path/file.ts:12)
  Why it matters: <reachable/visible/confusing behavior>
  Smallest cleanup: <implement/remove/document/issue>

### Probably Benign

- <candidate that looked suspicious but was superseded or test-only>

### Suggested Issues

- <one issue per actionable cleanup, with labels if the tracker supports them>

Bundled Resources

  • scripts/scan_litter.py: candidate scanner that emits Markdown or JSON.
  • references/signal-catalog.md: classification guide for deciding what is real litter versus normal code.