CLAUDE.md is the instruction file that Claude Code loads at the start of every session. It lives in the repo root or in .claude/. It tells the agent the stack, the commands, the file boundaries, and the rules. Every line is sent on every turn, so the file should be short and specific.

What should a CLAUDE.md contain?

Name the project and stack. List the build, test, and deploy commands. Name the folders the agent may and may not edit. State the style rules and the do-not rules. Skip general advice the model already knows. This reviewer flags the lines that cost tokens without changing behavior.

CLAUDE.md vs AGENTS.md: what is the difference?

CLAUDE.md is read by Claude Code. AGENTS.md is the open convention read by OpenAI Codex and many other agents. The content overlaps almost fully. Review the file here, then use the CLAUDE.md Converter to produce AGENTS.md, .cursorrules, Copilot instructions, or Windsurf rules.

How do I review a CLAUDE.md for token waste?

Paste the file into the reviewer and run the audit. It looks for vague rules, duplicate advice, missing commands, missing repo boundaries, and paragraphs that cost tokens every Claude Code turn without changing behavior.

What should a good CLAUDE.md include?

A strong CLAUDE.md usually names the project, stack, key commands, file boundaries, test/deploy workflow, style conventions, and concrete do/don't rules. Short, enforceable rules work better than generic advice.

Can this help with AGENTS.md or Cursor rules?

Yes. Review the rules first, then use the CLAUDE.md Converter to turn the cleaned instructions into AGENTS.md, .cursorrules, Copilot instructions, or Windsurf rules.

Does the tool rewrite my entire CLAUDE.md?

No. It returns a numbered audit and a few targeted rewrite suggestions so you can decide what to keep. That is safer than blindly replacing your repo instructions.

Is my CLAUDE.md stored?

No. The text is sent to MarkdownMe's capped Amazon Bedrock endpoint to generate the review, and only aggregate usage counters are logged.

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. Continue into AGENTS.md, Cursor, Copilot, and Windsurf formats with the related converter.

After the review, convert the file with the CLAUDE.md Converter for AGENTS.md or Cursor, or check your Claude fundamentals on the free Claude Certified Developer practice test.

AI-powered. Your CLAUDE.md is reviewed by Amazon Nova Micro. No login. Daily-capped to stay free.

Start with a common CLAUDE.md audit

Use an example shape, then replace it with your real repo instructions before reviewing.

CLAUDE.md reviewer guides

Use these examples when you need to cut token waste, add missing commands, tighten repo boundaries, or convert clean rules to other agent formats.

Token WasteFind vague, repeated, or unused CLAUDE.md instructions that cost tokens every Claude Code turn without improving behavior.CommandsCheck whether a CLAUDE.md includes the build, test, lint, dev server, deploy, and verification commands an agent needs before editing.BoundariesWrite CLAUDE.md rules that tell agents which files to edit, which files to avoid, and how to handle dirty worktrees safely.Convert After ReviewClean a CLAUDE.md first, then convert it into AGENTS.md, Cursor rules, GitHub Copilot instructions, or Windsurf rules.What BelongsDecide 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.Not FollowedDiagnose 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.Rules vs SkillsRoute each instruction to the right file: always-loaded CLAUDE.md, path-scoped .claude/rules/, or on-demand skills. One question decides it.Best PracticesRun your CLAUDE.md through eight checks: size, specificity, exact commands, anti-patterns, structure, contradictions, enforcement, and staleness.

Paste your CLAUDE.md

0 / 20,000 chars

Review & suggestions

What gets reviewed

• Missing sections (commands, file structure, conventions)
• Vague or generic instructions
• Over-long explanations that waste tokens per turn
• Missing behavioral rules (do X only, don't do Y)
• Inconsistencies and contradictions
• Returns numbered suggestions + 1-3 rewritten paragraphs

Today's picks

Three books for ticket and docs work. Check the current price before you buy.

Book · JiraAutomate Everyday Tasks in JiraNo-code automation rulesCheck price on Amazon ↗Book · ConfluenceImplementing Atlassian ConfluenceSpace and page structureCheck price on Amazon ↗Book · DocsHandbook of Technical Writing (10th ed.)Alphabetical reference formatCheck price on Amazon ↗

Amazon affiliate links. We earn a commission on qualifying purchases at no extra cost to you.

Was this rewrite usable?

Get new tools by email

We add tools most weeks. Give an email address and we send you the new ones, plus short notes on how to use them. No more than one email a week. Unsubscribe with one click.

Common Workflows

Use these patterns when you need this tool for a real publishing, documentation, or developer workflow.

Audit a bloated CLAUDE.md for token waste

Claude Code users whose project instructions have grown into vague, repetitive, or expensive context.

Paste the existing CLAUDE.md
Review missing sections, vague rules, and token-waste notes
Apply the suggested paragraph rewrites manually

A tighter agent instruction file that costs fewer tokens and gives clearer coding behavior.

Search intent: review CLAUDE.md for token waste

Improve agent behavior rules before a coding session

Teams using Claude Code, Codex, Cursor, or similar agents that need concrete project rules.

Paste the current agent instruction file
Look for missing commands, file boundaries, and do-not-do rules
Copy the strongest rewrite suggestions into the file

A more actionable CLAUDE.md that tells the agent how to work in the repo.

Search intent: improve Claude Code CLAUDE.md instructions

Review generated CLAUDE.md before committing

Developers who generated a first draft and want a second pass before checking it into the repo.

Generate or paste the CLAUDE.md draft
Run the reviewer
Keep only the concrete rules and remove generic filler

A commit-ready instruction file with fewer vague rules.

Search intent: CLAUDE.md reviewer for Claude Code

Convert cleaned Claude Code instructions to other agent formats

Developers who want one source of repo rules for Claude Code, Codex, Cursor, GitHub Copilot, and Windsurf.

Review the CLAUDE.md for vague or duplicate instructions
Apply the concrete fixes
Open the CLAUDE.md Converter and generate AGENTS.md, .cursorrules, or Copilot instructions

A cleaned instruction source that can be reused across multiple AI coding assistants.

Search intent: CLAUDE.md to AGENTS.md after review

Next useful step

Test your AI fundamentals next

If this tool is part of your AI or developer workflow, the free CloudYeti diagnostic shows which AWS AI concepts you already know and where you are weak.

Take the AI diagnostic

How to Use CLAUDE.md Reviewer

Open the CLAUDE.md Reviewer
Choose an example audit shape or paste your real CLAUDE.md
Run the AI review
Apply the numbered suggestions that match your repo
Convert the cleaned instructions to AGENTS.md, Cursor rules, Copilot instructions, or Windsurf rules if needed

Frequently Asked Questions

What is CLAUDE.md?: CLAUDE.md is the instruction file that Claude Code loads at the start of every session. It lives in the repo root or in .claude/. It tells the agent the stack, the commands, the file boundaries, and the rules. Every line is sent on every turn, so the file should be short and specific.
What should a CLAUDE.md contain?: Name the project and stack. List the build, test, and deploy commands. Name the folders the agent may and may not edit. State the style rules and the do-not rules. Skip general advice the model already knows. This reviewer flags the lines that cost tokens without changing behavior.
CLAUDE.md vs AGENTS.md: what is the difference?: CLAUDE.md is read by Claude Code. AGENTS.md is the open convention read by OpenAI Codex and many other agents. The content overlaps almost fully. Review the file here, then use the CLAUDE.md Converter to produce AGENTS.md, .cursorrules, Copilot instructions, or Windsurf rules.
How do I review a CLAUDE.md for token waste?: Paste the file into the reviewer and run the audit. It looks for vague rules, duplicate advice, missing commands, missing repo boundaries, and paragraphs that cost tokens every Claude Code turn without changing behavior.
What should a good CLAUDE.md include?: A strong CLAUDE.md usually names the project, stack, key commands, file boundaries, test/deploy workflow, style conventions, and concrete do/don't rules. Short, enforceable rules work better than generic advice.
Can this help with AGENTS.md or Cursor rules?: Yes. Review the rules first, then use the CLAUDE.md Converter to turn the cleaned instructions into AGENTS.md, .cursorrules, Copilot instructions, or Windsurf rules.
Does the tool rewrite my entire CLAUDE.md?: No. It returns a numbered audit and a few targeted rewrite suggestions so you can decide what to keep. That is safer than blindly replacing your repo instructions.
Is my CLAUDE.md stored?: No. The text is sent to MarkdownMe's capped Amazon Bedrock endpoint to generate the review, and only aggregate usage counters are logged.

Related Tools

CLAUDE.md Converter

Convert agent configs between CLAUDE.md, AGENTS.md, .cursorrules, .github/copilot-instructions.md, and .windsurfrules. One file, every format.

CLAUDE.md Generator

Describe your project in 1-2 sentences and get a structured CLAUDE.md. AI-powered by Amazon Nova Micro. Free, no login.

Improve My Prompt

Paste any LLM prompt and get specific, actionable fixes against 12 best-practice rules from Anthropic, OpenAI, and Google. Free, no login.

AI Prompt Optimizer (Rewrites Your Prompt)

Paste a rough prompt and get a structured, production-grade rewrite — Role, Context, Task, Constraints, Output Format, and Examples — tuned to best practices from Anthropic, OpenAI, and Google. Powered by Amazon Nova.

About CLAUDE.md Reviewer

Paste your existing CLAUDE.md and get specific, numbered suggestions for vague rules, missing commands, repo boundaries, and token waste. This AI-backed tool sends your input to MarkdownMe's capped Amazon Bedrock endpoint to generate the result. No account is required. MarkdownMe provides free, fast, and reliable Markdown tools for developers, writers, and anyone who works with Markdown.