Claude Code and CLAUDE.md tools
Your CLAUDE.md file shapes every Claude Code session. These tools help you review it for vague rules and token waste, generate one from scratch, and convert it to other agent formats. Free, no signup.
CLAUDE.md Reviewer
Claude Code reads your CLAUDE.md every session, so every vague or unused line costs tokens and weakens behavior. Paste your CLAUDE.md and get a numbered audit: missing commands, missing file boundaries, generic rules, over-long explanations, contradictions, and 1-3 rewritten paragraphs you can drop in directly. Need to create one from scratch? Use the CLAUDE.md generator. Continue into AGENTS.md, Cursor, Copilot, and Windsurf formats with the related converter. Reviewing an AGENTS.md or a Claude skill instead? Use the AGENTS.md Reviewer or the SKILL.md Reviewer.
AGENTS.md Reviewer
AGENTS.md is the open instruction file that OpenAI Codex reads, and Cursor, GitHub Copilot, Google Jules, Amp, and other coding agents read it too. Agents re-read it on every task, so a missing test command or a vague rule costs you on every run. Paste your AGENTS.md and get a numbered audit: setup, build, test, and lint commands that are missing or not runnable, rules that belong in a nested AGENTS.md inside a monorepo package, missing PR and commit conventions, files agents must not touch, vague lines, token bloat, and Claude Code-only syntax (@imports, slash commands, hooks) that other agents ignore. The review ends with 1-3 rewritten paragraphs you can paste in. Starting from a Claude Code file? Review it with the CLAUDE.md Reviewer, then produce AGENTS.md with the CLAUDE.md Converter. Writing a Claude Agent Skill? Use the SKILL.md Reviewer.
SKILL.md Reviewer
A Claude Agent Skill is a folder with a SKILL.md file. Claude loads only the name and description from its YAML frontmatter until the skill triggers, so a vague description means the skill never runs, and a bloated body wastes context when it does. Paste your SKILL.md and get a numbered audit: a valid kebab-case name, a description that says what the skill does and when to use it with concrete trigger phrases, imperative steps instead of narration, long reference material moved into separate files referenced by path, concrete examples, scripts referenced by path, and a body under about 500 lines. The review ends with a rewritten description: line and 1-2 rewritten sections. For always-loaded project instructions, use the CLAUDE.md Reviewer. For the cross-agent instruction file, use the AGENTS.md Reviewer.
CLAUDE.md Generator
Type a short description of your project — stack, deploy target, key dependencies — and get back a polished CLAUDE.md with all six standard sections (overview, tech, file structure, commands, workflows, behavioral rules). AI-powered. Daily-capped to stay free and fast.
CLAUDE.md Converter
Paste your CLAUDE.md (or AGENTS.md, or raw rules) and instantly convert it into the format your other agent tool expects. Handles Claude Code, OpenAI Codex, Cursor, GitHub Copilot, and Windsurf. Preserves sections for markdown targets, extracts imperative rules for flat rule-files. Download or copy. No upload, all client-side.
CLAUDE.md how-to guides
Step-by-step answers for writing, reviewing, and converting CLAUDE.md files.
Find vague, repeated, or unused CLAUDE.md instructions that cost tokens every Claude Code turn without improving behavior.
CLAUDE.md Missing Commands ChecklistCheck whether a CLAUDE.md includes the build, test, lint, dev server, deploy, and verification commands an agent needs before editing.
CLAUDE.md Repo Boundaries and RulesWrite CLAUDE.md rules that tell agents which files to edit, which files to avoid, and how to handle dirty worktrees safely.
Review CLAUDE.md Before Converting to AGENTS.mdClean a CLAUDE.md first, then convert it into AGENTS.md, Cursor rules, GitHub Copilot instructions, or Windsurf rules.
What Should Go in a CLAUDE.md FileDecide what belongs in CLAUDE.md and what belongs in the README, a path-scoped rule, or a skill. One test per line, applied to a real file.
Claude Code Is Not Following CLAUDE.md — How to Fix ItDiagnose why Claude Code ignores your CLAUDE.md: the file never loaded, the rule is vague, two files contradict, or the rule needs a hook instead.
CLAUDE.md vs .claude/rules vs Skills — Which File Owns WhatRoute each instruction to the right file: always-loaded CLAUDE.md, path-scoped .claude/rules/, or on-demand skills. One question decides it.
CLAUDE.md Best Practices Checklist — Audit Your File Against 8 RulesRun your CLAUDE.md through eight checks: size, specificity, exact commands, anti-patterns, structure, contradictions, enforcement, and staleness.
CLAUDE.md FAQ
What is CLAUDE.md?
CLAUDE.md is a project instruction file that lives in a repository root and tells Claude Code what your project needs: build commands, testing rules, file boundaries, and conventions. Claude Code loads it in every session, so it shapes every interaction.
What should a CLAUDE.md contain?
Include the commands Claude needs: build, test, lint, dev server, and deploy. Add your project layout, naming conventions that differ from tool defaults, and two real anti-patterns your team has hit. Skip generic advice like write good code and follow existing patterns. Keep it under 200 lines.
CLAUDE.md vs AGENTS.md vs .cursorrules: what is the difference?
CLAUDE.md is for Claude Code. AGENTS.md is the shared format that Codex and several other agents read. .cursorrules is for Cursor. They all tell an agent the same kinds of facts about your project, but they load in different editors. Many teams maintain one clean CLAUDE.md and generate the other formats from it.
Does a long CLAUDE.md cost tokens every turn?
Yes. Claude Code loads your CLAUDE.md into every session, so every line consumes tokens on every turn. That is why cutting vague rules and stale notes pays off: a 50-line file costs less than a 200-line file even when they say the same thing.
Working with Cursor instead? The CLAUDE.md Converter generates .cursorrules from your CLAUDE.md. For Copilot, Windsurf, or Codex, convert to those formats too.