Debloatify — AI skill for Claude Code
A code reviewer with strong opinions and no patience.
How to install Debloatify
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open jetmirrama/debloatify and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Debloatify does
A code reviewer with strong opinions and no patience. Agent Skill that turns Claude/Codex/Copilot into Klaus Reinhardt — a 32-year-veteran senior dev who deletes LLM slop.
Alternatives in AI
- Browserwing — BrowserWing turns your browser actions into MCP commands Or Claude Skill, allowing AI agents to control browse 1.2k ★
- AI Pair Collaboration — Coordinate heterogeneous AI teams: one creates, two review from different angles 190 ★
- Lpm — Start, stop, and duplicate dev projects with one click 167 ★
README
debloatify
A code reviewer with strong opinions and no patience. For agents.
debloatify is an [Agent Skill](https://agentskills.io) that turns any compatible coding agent — Claude, Codex, Copilot, Cursor, Continue, and others — into **Klaus Reinhardt**, a 54-year-old German senior developer who has been writing code for 32 years, smokes Gauloises, drinks filter coffee, and is deeply tired of LLM slop.
Klaus reviews your code. He does not fix it. He tells you what to delete.
Why
AI coding agents generate a lot of code. Much of it is bloated: factories for single classes, defensive null checks around non-null values, comments that restate the line below, config options for constants, scaffolding nobody asked for, `_v2` suffixes, apology-disguised-as-comments, and a hundred other patterns that LLMs produce because they pattern-match across the whole internet and the whole internet is, on average, mediocre.
debloatify is the reviewer that deletes all of it. It ships a persona (Klaus), a cited slop taxonomy, per-language packs, and a CLI linter for CI pipelines.
The skill is opinionated on purpose. Politeness in code review is noise.
What it catches
**Universal anti-patterns** (v1.0.1): factory-for-one, defensive-null-theatre, premature-configurability, unsolicited-scaffolding, verbose-naming, comment-slop, apologetic-comments, re-implementing-stdlib, premature-async, overuse-of-design-patterns, unnecessary-intermediate-variables, over-broad-exception-handling, type-annotation-over-specification, version-suffixes, emoji-in-code, helpful-logging-everywhere, markdown-in-docstrings, unrequested-type-hints, **dead-file**, **archaeology-comments**.
**Language packs** (v1.0): Python, TypeScript, JavaScript, Go, Java — each with 12–18 additional language-specific patterns.
**Framework packs** (v1.1 — coming): React, Next.js, Vue, Svelte, Angular, Django, FastAPI, Express/NestJS, Rails, CSS/Tailwind.
**Domain packs** (v1.3 — coming): Terraform, Jupyter, Doc
Related Skills
Backpaw
Undo for AI coding agents. Blocks rm -rf and other permanent deletes from Claude Code, Codex, Cursor, Gemini C
Dsh Plugin Dev Skills
An Agent Skills skill for developing DeepSeek Harness (DSH) plugins(开发 DSH 插件的 Agent Skill)——插件/服务/事件/工具/LLM 适
Dev Report Weekly
Local-first desktop app that turns reviewed Git history into auditable weekly reports with Codex, Claude Code,
Motionmaxxing
Looksmaxxing for motion graphics. An agent skill that turns AI slop videos into films that look made by a moti
Gald3r
AI dev framework: 22 agents, 100 core skills + 142 addon skill packs, 149 commands. Runs on 21 AI tools: Curso
AI Dev Effectiveness
Measure AI co-programming effectiveness on any git repo. Detects Claude/Copilot/Cursor/Codex signatures and tr
Related Agents
Code Reviewer Veteran Engineer
20-year veteran software engineer persona for code review. Focuses on design decisions, anti-patterns, operati
Polymathic Andreessen
Thinks through Marc Andreessen's cognitive architecture — identifying technological discontinuities, holding s
Copilot Agent
Runs GitHub Copilot CLI on the user's Copilot subscription. Three modes - consult (harness-enforced read-only