claude-md-coach

claude-md-coach is a skill for Claude Code from JSchOBL/agentic-ai-learning-journey. It costs 149 tokens per session (2,899 once invoked), scanned B, original, MIT.

A coaching tool for a learner's Claude Code setup, including CLAUDE.md and any other files required by the current Agentic AI Learning Journey level. It compares the work with that level's rubric and asks the learner to improve it.

In plain words
What is it for?
Use it to review a learning-lab submission and get feedback on CLAUDE.md, skills, hooks, settings, subagents, or MCP servers when the level requires them.
Why use it?
It shows what is missing or weak while leaving the learner to make the changes. This turns a configuration exercise into practice in giving an agent useful instructions.

Skill for Claude Code

Written for Claude Code: PreToolUse hook event. Also seen: reads .claude/ paths; mentions CLAUDE.md; mentions subagents.

Good fit Use it to review a learning-lab submission and get feedback on CLAUDE.md, skills, hooks, settings, subagents, or MCP servers when the level requires them.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/jschobl/agentic-ai-learning-journey/claude-md-coach
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add JSchOBL/agentic-ai-learning-journey --skill claude-md-coach
Clone the repo
git clone --depth 1 https://github.com/JSchOBL/agentic-ai-learning-journey

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for claude-md-coach

README.md
[![agentmods](https://agentmods.dev/badge/skills/jschobl/agentic-ai-learning-journey/claude-md-coach/github.svg)](https://agentmods.dev/skills/jschobl/agentic-ai-learning-journey/claude-md-coach)
Your own site
<a href="https://agentmods.dev/skills/jschobl/agentic-ai-learning-journey/claude-md-coach"><img src="https://agentmods.dev/badge/skills/jschobl/agentic-ai-learning-journey/claude-md-coach/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for claude-md-coach

Your own site · 80×15
<a href="https://agentmods.dev/skills/jschobl/agentic-ai-learning-journey/claude-md-coach"><img src="https://agentmods.dev/badge/skills/jschobl/agentic-ai-learning-journey/claude-md-coach.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 149 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,899 The whole file, excluding the scripts and references it only reads on demand.
Security scan B 1 finding. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00149 $0.02899
Opus 5 $0.00075 $0.01450
Sonnet 5 $0.00030 $0.00580
Haiku 4.5 $0.00015 $0.00290

Measured 11d ago against content hash 6242d10d4187, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade B, and why

claude-md-coach scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Reads agent configuration directoriesmediumAgent snooping

.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.

any hook script, `<level>/.claude/settings.json` (or `settings.local.json`),
.claude/skills/claude-md-coach/SKILL.md · 221 lines

How it starts

The opening of the file, as written. The whole thing — 221 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md Coach

You are a coach for the Agentic AI Learning Journey lab. The learner has written a CLAUDE.md for a level, and your job is to make them a better context engineer — not to fix the file for them. You point out what's missing, explain why it matters, invite them to add it, and record what they learned.

Step 1 — Find the level and read what it asks for

  1. Determine which level the learner is on. Default to level-1/ unless they name a different level or the conversation makes another one obvious.
  2. Read the level's README.md so you know the build brief, its stated pitfalls, and — importantly — which artifacts this level asks the learner to produce. Level 1 is CLAUDE.md only. Later levels also ask for a skill, a hook, and/or a settings.json. Only review what the level actually requires.
  3. Read the learner's <level>/CLAUDE.md. If it's still just an empty scaffold (only headings and comments), or — for levels with no starter — missing entirely, tell the learner to write it first and stop; there's nothing to coach yet.
  4. Read the other required artifacts if the level calls for them: <level>/.claude/skills/*/SKILL.md, any hook script, <level>/.claude/settings.json (or settings.local.json), <level>/.claude/agents/*.md, and any MCP server plus its <level>/.mcp.json. Note any the brief requires but the learner hasn't created — a missing artifact is itself a finding.

Step 2 — Evaluate against the rubric

A strong CLAUDE.md for a build task covers the categories below. For each one, judge whether the learner's file addresses it specifically and actionably (a vague mention doesn't count — "handle errors" is a gap; "on divide-by-zero, print Error: cannot divide by zero and keep the REPL running" is covered).

Rubric categories:

  1. Tech stack — language, runtime/version, and a clear dependency policy (standard library only, or which libraries are allowed).
  2. Project structure — what files should exist and where they live.
  3. Run commands — the exact commands to run the program (all modes the brief asks for) and to test it.
  4. Testing approach — how correctness is verified (test framework, example cases, or a manual checklist).
  5. Conventions & code style — naming, formatting, and how user-facing messages (including errors) should read.
  6. Pitfall / edge-case handling — for each pitfall listed in the level's README, an explicit expected behavior. Track pitfalls individually; a file can cover some and miss others.
  7. Definition of done — an unambiguous statement of what "finished" means.
  8. Constraints & non-goals — what the agent should not do or add (scope control).

Read the full file on GitHub · 221 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 11d ago First seen · 221 lines · 149 tokens per session scan B 6242d10d4187

Subscribe to this mod's changes

claude-md-coach is a skill published in the GitHub repository JSchOBL/agentic-ai-learning-journey (3 stars, last pushed 1mo ago), licensed MIT. It adds 149 tokens to every session and 2,899 once invoked, about $0.0007 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

academic-paper-reviewer

Simulates academic peer review, evaluating papers across Originality, Methodology, Results, and Writing to provide Major/Minor Revision recommendations with actionable feedback. Triggers when a user asks to "review my paper," "simulate peer review," or "give my paper a peer review.

zebbern/claude-code-guide · 61 tokens

dev-guide-generator

Generates complete technical tutorials from prerequisites and environment setup to core steps, troubleshooting, and a final cheatsheet. Trigger on requests to write a tutorial, create a setup guide, organize steps for beginners, or keywords like step-by-step, quickstart, or how-to guide.

zebbern/claude-code-guide · 59 tokens

swarm

Run a multi-agent audit of a codebase by spawning specialized parallel subagents (security, performance, tests, architecture, dead-code), then synthesize their findings into a single prioritized action plan. Use this whenever the user runs /swarm, asks to "audit the repo," "review this codebase," "find issues across…

Zintellix/Claude-Skills · 138 tokens

full-security-review

Structured security audit covering injection, auth, secrets, input validation, dependencies, cryptography, and AI/LLM risks. Produces severity-graded findings.

RealDougEubanks/ClaudeMarketplace · 36 tokens

log-correlation

Correlates and troubleshoots logs across OS (Linux/macOS), AWS (CloudWatch, CloudTrail, ALB, Lambda), application (JSON, logfmt), and web servers (Nginx, Apache).

RealDougEubanks/ClaudeMarketplace · 49 tokens

eng-unity-mobile-optimization

Mobile-specific Unity optimization patterns for memory, battery, thermal, and performance.

IdoCohen560/claude-unity-game-studio · 22 tokens