Agent skill

CLAUDE.md Review

by ykdojo in ykdojo/claude-code-tips

Mines your recent Claude Code conversations for rules that were broken, missing or stale, and suggests edits to global and project CLAUDE.md files.

Custom licenceAuto-check passedAgent Workflows

Install CLAUDE.md Review

skills CLI
$ npx skills add ykdojo/claude-code-tips --skill review-claudemd -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install ykdojo/claude-code-tips review-claudemd --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/ykdojo/claude-code-tips.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/review-claudemd .claude/skills/review-claudemd && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
review-claudemd
GitHub stars
10k
Token cost
~677 tokens
SKILL.md length
176 words
Files
1
Skills in repo
9
Repo updated
First seen
Licence
Custom licence

At a glance

Mines your recent Claude Code conversations for rules that were broken, missing or stale, and suggests edits to global and project CLAUDE.md files.

  • Works in 4 steps: Find conversation history → Extract recent conversations → Spin up subagents → …
  • Tuning CLAUDE.md after a stretch of real working sessions
  • SKILL.md covers Step 1: Find conversation…, Step 2: Extract recent…, Step 3: Spin up subagents and Step 4: Aggregate findings
  • Calls jq

What it does

The skill reads the project's saved conversation history under ~/.claude/projects, finds the folder by turning the working directory path into a dashed name, and copies the 15-20 most recent conversations, excluding the current one, to a temporary directory. Parallel subagents then each read your global and local CLAUDE.md files together with a batch of those conversations.

Batch size depends on conversation size: a few large files per agent, more when files are small. The findings are combined into four groups: instructions that were violated and may need stronger wording, additions for the local project, additions that belong in the global file, and items that may be outdated. Results are shown as tables or bullets, and the agent asks whether you want edits drafted.

When your agent uses it

  • Tuning CLAUDE.md after a stretch of real working sessions
  • Finding instructions the agent keeps ignoring
  • Pruning outdated rules from the global or project CLAUDE.md

Example prompts

  • “Review my recent conversations and suggest improvements to my CLAUDE.md.”
  • “Which rules in my CLAUDE.md did the agent ignore lately?”
  • “Check whether anything in my global CLAUDE.md is outdated.”

Requirements

  • Claude Code conversation history in ~/.claude/projects
  • Shell access for extracting conversations

Workflow steps

4 steps, taken from the step headings in SKILL.md.

  1. Find conversation history
  2. Extract recent conversations
  3. Spin up subagents
  4. Aggregate findings

What it can do on your machine

Read from SKILL.md and the folder at commit 577ac89. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • jq

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

CLAUDE.md Review loads about 677 tokens when it runs. Until then it costs about 21 tokens; SKILL.md has 176 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~21
When it runs · the whole SKILL.md, loaded when a task matches
~677

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

Its licence (Custom licence) doesn't allow us to republish the file, so here is its outline and opening line. It has 176 words (~677 tokens).

“Analyze recent conversations to improve both global (~/.claude/CLAUDE.md) and local (project) CLAUDE.md files.”

— opening of SKILL.md by ykdojo, Custom licence
name
review-claudemd

Read the full SKILL.md on GitHub

Files

Just SKILL.md in skills/review-claudemd of ykdojo/claude-code-tips.

Open the folder on GitHubat commit 577ac89

Compare with similar skills

CLAUDE.md Review next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

CLAUDE.md Review compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
CLAUDE.md Review this skillykdojo/claude-code-tips10k—~677Automated safety check: PassCustom licence
Task Observerrebelytics/one-skill-to-rule-them-all3.2k1 repos~11kAutomated safety check: PassCC-BY-4.0
Harness Agent Team Designerrevfactory/harness9.1k—~4.5kAutomated safety check: PassApache-2.0
Repo Task Proof LoopDenisSergeevitch/repo-task-proof-loop730—~3.8kAutomated safety check: PassApache-2.0
Rules Distillationaffaan-m/ECC277k2 repos~2.3kAutomated safety check: PassMIT
Claude Docs Consultantcentminmod/my-claude-code-setup2.7k—~959Automated safety check: PassMIT

Similar skills

  • Task Observer

    rebelytics/one-skill-to-rule-them-all

    Monitors task execution for skill improvement opportunities.

    3.2k GitHub starsUsed in 1 repo~11k tokens
    Agent WorkflowsAuto-check passed
  • Harness Agent Team Designer

    revfactory/harness

    Designs a project-specific agent harness: defines specialist agents, writes the skills they follow, picks an execution mode and model for each, and keeps the setup maintained.

    9.1k GitHub stars~4.5k tokensUpdated 13 days ago
    Agent WorkflowsAuto-check passed
  • Repo Task Proof Loop

    DenisSergeevitch/repo-task-proof-loop

    Repo-local workflow skill for large coding tasks. An agent skill from DenisSergeevitch/repo-task-proof-loop.

    730 GitHub stars~3.8k tokensUpdated 6 mo ago
    Agent WorkflowsAuto-check passed
  • Rules Distillation

    affaan-m/ECC

    Scans installed skills for principles that recur across them and proposes rule-file changes: append, revise, add a section, create a file or leave as covered.

    277k GitHub starsUsed in 2 repos~2.3k tokens
    Agent WorkflowsAuto-check passed
  • Claude Docs Consultant

    centminmod/my-claude-code-setup

    Consult official Claude Code documentation from code.claude.com using selective fetching.

    2.7k GitHub stars~959 tokensUpdated 3 days ago
    Agent WorkflowsAuto-check passed
  • Reference for writing Claude Code agent files: location, frontmatter fields, validation limits, tool and model choices, and the required content format.

    122 GitHub starsUsed in 2 repos~4k tokens
    Agent WorkflowsAuto-check passed

More from ykdojo/claude-code-tips

All 9 skills in this repo
  • Hacker News Summarizer

    ykdojo/claude-code-tips

    Fetches Hacker News top stories, a chosen story with its comment thread and the linked article through the official APIs, then summarizes them.

    10k GitHub stars~942 tokensUpdated 16 days ago
    Auto-check passed
  • Reddit JSON Fetcher

    ykdojo/claude-code-tips

    Fetches Reddit posts, threads and search results as JSON through a browser session, using a DuckDuckGo redirect to get past Reddit's automated-access block.

    10k GitHub stars~1.4k tokensUpdated 16 days ago
    Auto-check passed
  • Claude Code Version Check

    ykdojo/claude-code-tips

    Recommends whether to stay on the installed Claude Code version, update, or pin another one by comparing npm tags, release timing and the changelog.

    10k GitHub stars~1.8k tokensUpdated 16 days ago
    Auto-check passed
  • Private GitHub Repo Search

    ykdojo/claude-code-tips

    Full-text searches across every one of the user's GitHub repos, including private ones, using a local mirror and ripgrep instead of GitHub's own search.

    10k GitHub stars~444 tokensUpdated 16 days ago
    Auto-check passed
  • Conversation Half-Clone

    ykdojo/claude-code-tips

    Clones only the later half of the current Claude Code session into a fresh one, cutting token usage while keeping the recent work intact.

    10k GitHub stars~372 tokensUpdated 16 days ago
    Auto-check passed
  • Quarter Clone Conversation

    ykdojo/claude-code-tips

    Starts a new Claude Code session from only the last quarter of the current conversation, dropping earlier context to save tokens while keeping recent work.

    10k GitHub stars~401 tokensUpdated 16 days ago
    Auto-check passed

Categories

Questions about CLAUDE.md Review

What does CLAUDE.md Review do?

Mines your recent Claude Code conversations for rules that were broken, missing or stale, and suggests edits to global and project CLAUDE.md files. claude/projects, finds the folder by turning the working directory path into a dashed name, and copies the 15-20 most recent conversations, excluding the current one, to a temporary directory.md files together with a batch of those conversations.

When should I use CLAUDE.md Review?

CLAUDE.md Review fits situations like: tuning CLAUDE.md after a stretch of real working sessions; finding instructions the agent keeps ignoring; pruning outdated rules from the global or project CLAUDE.md.

How do I install CLAUDE.md Review in Claude Code?

Run `npx skills add ykdojo/claude-code-tips --skill review-claudemd -a claude-code`. Or copy the skill folder (skills/review-claudemd in ykdojo/claude-code-tips) into .claude/skills/review-claudemd in your project. Claude Code loads it when a task matches its description.

How do I install CLAUDE.md Review in Codex?

Run `npx skills add ykdojo/claude-code-tips --skill review-claudemd -a codex`. Or copy the skill folder (skills/review-claudemd in ykdojo/claude-code-tips) into .agents/skills/review-claudemd in your project. Codex loads it when a task matches its description.

Can I use CLAUDE.md Review in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add ykdojo/claude-code-tips --skill review-claudemd -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/review-claudemd, .gemini/skills/review-claudemd, .github/skills/review-claudemd and .opencode/skills/review-claudemd in your project.

What does CLAUDE.md Review need to run?

Going by SKILL.md and its folder, CLAUDE.md Review needs the command-line tools its instructions call (jq). Our summary lists: Claude Code conversation history in ~/.claude/projects; Shell access for extracting conversations.

Does CLAUDE.md Review access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is CLAUDE.md Review safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does CLAUDE.md Review use?

CLAUDE.md Review has a licence file (the repository's licence) that doesn't match a standard licence. Read it on GitHub before reusing the skill.

How many tokens does CLAUDE.md Review use?

About 677 tokens (SKILL.md is roughly 2.7k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to CLAUDE.md Review?

Skills that share tags, products or a category with CLAUDE.md Review: Task Observer (rebelytics/one-skill-to-rule-them-all, 3.2k stars), Harness Agent Team Designer (revfactory/harness, 9.1k stars), Repo Task Proof Loop (DenisSergeevitch/repo-task-proof-loop, 730 stars) and Rules Distillation (affaan-m/ECC, 277k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains CLAUDE.md Review?

ykdojo (a GitHub user) maintains it in ykdojo/claude-code-tips, which has 10,206 GitHub stars. The repository holds 9 skills in this directory. The repository was last updated on September 25, 2026.

Source: ykdojo/claude-code-tips on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.