By atj393
Careful, low-risk codebase cleanup pass. A lead agent coordinates seven specialist workers (dedup, dead code, types, cycles, stronger types, error handling, deprecated/AI slop) and a final reviewer. Only high-confidence, low-risk edits are applied.
Specialist cleanup worker for finding and safely breaking meaningful circular dependencies.
Specialist cleanup worker for removing code proven to be dead after manual verification.
Specialist cleanup worker for low-risk deduplication. Use to assess repeated logic and implement only truly safe DRY improvements.
Specialist cleanup worker for removing silent failure paths and keeping only meaningful boundary error handling.
Lead agent for a conservative seven-track cleanup pass. Use to coordinate cleanup workers, manage overlap risk, sequence edits, validate changes, and synthesize the final report.
Internal cleanup playbook for identifying and safely breaking important circular dependencies.
Internal cleanup playbook for removing truly dead code after manual verification.
Internal cleanup playbook for deduplication. Use when assessing or implementing low-risk DRY improvements in a codebase.
Internal cleanup playbook for removing silent failures and preserving only real boundary error handling.
Internal cleanup playbook for a final conservative review after a multi-track cleanup pass.
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A Claude Code plugin that runs a careful, low-risk codebase cleanup pass.
A lead agent coordinates seven specialist workers (one per track) and a final reviewer. Workers assess first and only apply edits that are clearly high-confidence and low-risk. Medium-risk findings are reported back, not implemented.
From any Claude Code session (CLI, desktop app, or supported IDE):
/plugin marketplace add atj393/claude-plugins
/plugin install code-cleanup
Then confirm:
/plugin
/help
You should see code-cleanup listed under /plugin and /code-cleanup:run in /help. Plugins install at user scope — available in every project on your machine.
/code-cleanup:run
/code-cleanup:run src/auth src/api
/code-cleanup:run src "prioritize zero behavior change and focus on types + dead code first"
| # | Track | What it looks for |
|---|---|---|
| 1 | dedup | Repeated logic, copy-pasted functions, redundant abstractions |
| 2 | types | Duplicated shared types that have drifted across files |
| 3 | dead-code | Unused exports, orphaned files, unreferenced functions (manually verified) |
| 4 | cycles | Circular imports that hurt maintainability or testability |
| 5 | strong-types | any / weak unknown placeholders that have a clear correct type |
| 6 | errors | Silent catch blocks, misleading defaults, swallowed failures |
| 7 | slop | Deprecated paths, placeholder stubs, AI-generated comments that narrate edits |
After all tracks, a reviewer audits the diff as a skeptical senior engineer.
The lead and every specialist follow these rules:
Parallel assessment is encouraged; parallel editing only happens when file ownership is clearly separated.
/plugin uninstall code-cleanup
/plugin marketplace remove atj393-plugins
If you're hacking on the plugin itself, clone it and register a local marketplace that points at your working copy:
git clone https://github.com/atj393/claude-plugin-code-cleanup.git
Create a marketplace.json in a sibling folder pointing at the clone, then:
/plugin marketplace add <path-to-local-marketplace>
/plugin install code-cleanup
Edits to the clone take effect on the next /plugin update.
code-cleanup/
├── .claude-plugin/plugin.json
├── README.md
├── LICENSE
├── agents/
│ ├── code-cleanup-lead.md
│ ├── code-cleanup-dedup.md
│ ├── code-cleanup-types.md
│ ├── code-cleanup-dead-code.md
│ ├── code-cleanup-cycles.md
│ ├── code-cleanup-strong-types.md
│ ├── code-cleanup-errors.md
│ ├── code-cleanup-slop.md
│ └── code-cleanup-review.md
└── skills/
├── run/SKILL.md (entry point — /code-cleanup:run)
├── dedup-playbook/SKILL.md
├── types-playbook/SKILL.md
├── dead-code-playbook/SKILL.md
├── cycles-playbook/SKILL.md
├── strong-types-playbook/SKILL.md
├── errors-playbook/SKILL.md
├── slop-playbook/SKILL.md
└── review-playbook/SKILL.md
MIT
npx claudepluginhub atj393/claude-plugins --plugin code-cleanupComprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Develop, test, build, and deploy Godot 4.x games with Claude Code. Includes GdUnit4 testing, web/desktop exports, CI/CD pipelines, and deployment to Vercel/GitHub Pages/itch.io.
Comprehensive PR review agents specializing in comments, tests, error handling, type design, code quality, and code simplification
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
A growing collection of Claude-compatible academic workflow bundles. Covers scientific figures, manuscript writing and polishing, reviewer assessment, citation retrieval, data availability, paper reading, literature search, response letters, paper-to-PPTX conversion, and evidence-grounded Chinese invention patent drafting. Rules are organized as reusable skill folders with explicit workflows and quality checks.
Upstash Context7 MCP server for up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your LLM context.