Claude MD Generator
Scans a codebase and drafts a tight CLAUDE.md of only non-obvious commands and gotchas
Test report
- Verdict
- Works with setup
- Score
- Tested
- Jul 21, 2026
- Environment
- Claude Code 2.x (agent harness)
- Upstream re-checked
- Aug 10, 2026 · 3ce28f7
Fetched SKILL.md + all referenced files (section-guide.md, .env.example, evals.json all HTTP 200); frontmatter parses with name+description. Step 6 hard-calls the Gemini API with $GEMINI_API_KEY (POST to generativelanguage.googleapis.com), which I did not have, so I executed its drafting rules directly in place of the API. Built a synthetic billing-svc project (vitest, drizzle, path aliases, a @generated file, env-ordering dependency) and produced two CLAUDE.md files: baseline (49 lines, redundant Tech Stack + npm install + inline snippet + vague Code Style, no Gotchas, missed the generated file, the alias values, and the DATABASE_URL-before-tests ordering) vs skill-rules (26 lines, concrete @billing/@shared aliases, flagged src/schema.gen.ts:1 as generated do-not-edit, captured the test env ordering, no filler). Verdict setup because a plain install genuinely needs a Gemini API key to run Step 6 as written. Docs note: SKILL.md says under 100 lines, README says under 200.
Scored on four weighted criteria — install, triggering, output vs. baseline, docs. How scoring works
- Installs cleanly 5/5
- Triggers reliably 5/5
- Output vs. baseline 7/10
- Docs & honesty 4/5
What Claude MD Generator does
Scans a project (package.json, tsconfig, Makefile, .env.example, generated files) and drafts a focused CLAUDE.md or AGENTS context file containing only the non-obvious build commands, path aliases, and gotchas Claude cannot infer from the code. Triggers when the user asks to generate, update, or audit a project's CLAUDE.md/AGENTS file. Its Step 6 draft-generation step calls the Google Gemini API and requires a GEMINI_API_KEY.
How to install Claude MD Generator
git clone --depth 1 https://github.com/Varnan-Tech/opendirectory.git /tmp/claude-md-generator-src
mkdir -p ~/.claude/skills
cp -R /tmp/claude-md-generator-src/skills/claude-md-generator ~/.claude/skills/claude-md-generator
# Requires a Gemini API key for Step 6 (draft generation):
# export GEMINI_API_KEY=... (get one at aistudio.google.com)
# The skill POSTs the compiled project analysis to generativelanguage.googleapis.com.
# Alt install (npx CLI): npx "@opendirectory.dev/skills" install claude-md-generator --target claude
# Alt install (marketplace): /plugin marketplace add Varnan-Tech/opendirectory ; /plugin install opendirectory-gtm-skills@opendirectory-marketplace
Skills live in ~/.claude/skills/ (global) or .claude/skills/
(per-project). Restart Claude Code after installing.
Commands — how to trigger Claude MD Generator
-
/claude-md-generatorScans a codebase and drafts a tight CLAUDE.md of only non-obvious commands and gotchas
It also activates on plain-language prompts like these:
-
Generate a CLAUDE.md for this repository -
Update AGENTS.md from a fresh codebase scan -
Write a concise context file with our build commands
Frequently asked questions
- Is the Claude MD Generator skill free?
- Yes. The skill itself is free from Varnan-Tech/opendirectory. SkillProof publishes the install command and an independent test verdict at no cost.
- Does Claude MD Generator work with Claude Code?
- We tested it with Claude Code 2.x (agent harness) on Jul 21, 2026. Verdict: Works with setup. Fetched SKILL.md + all referenced files (section-guide.md, .env.example, evals.json all HTTP 200); frontmatter parses with name+description. Step 6 hard-calls the Gemini API with $GEMINI_API_KEY (POST to generativelanguage.googleapis.com), which I did not have, so I executed its drafting rules directly in place of the API. Built a synthetic billing-svc project (vitest, drizzle, path aliases, a @generated file, env-ordering dependency) and produced two CLAUDE.md files: baseline (49 lines, redundant Tech Stack + npm install + inline snippet + vague Code Style, no Gotchas, missed the generated file, the alias values, and the DATABASE_URL-before-tests ordering) vs skill-rules (26 lines, concrete @billing/@shared aliases, flagged src/schema.gen.ts:1 as generated do-not-edit, captured the test env ordering, no filler). Verdict setup because a plain install genuinely needs a Gemini API key to run Step 6 as written. Docs note: SKILL.md says under 100 lines, README says under 200.
- What is the Claude MD Generator SkillProof Score?
- 8.4/10 — installs cleanly 5/5, triggers reliably 5/5, output vs. baseline 7/10, docs & honesty 4/5.
- How do I install Claude MD Generator?
- Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.
- Can I use Claude MD Generator with Cursor, Copilot, Gemini CLI, Codex or other AI tools?
- The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.