Manage project documentation for Claude Code workflows including context files, READMEs, and agent instructions. This skill provides structured automation for maintaining accurate, up-to-date documentation that aligns with actual codebase structure and functionality. Use this skill when initializing new projects, updating existing documentation, or ensuring context files accurately reflect current code.
Works with
AI-first code editor with Composer
Before installing skills in Cursor, ensure your development environment meets these requirements:
node --versionmd-docsExecute the skills CLI command in your project's root directory to begin installation:
Fetches md-docs from paulrberg/agent-skills and configures it for Cursor.
The CLI shows a list of agents. Use arrow keys and space to select Cursor:
Confirm successful installation by checking the skill directory location:
Restart Cursor to activate md-docs. Access via /md-docs in your agent's command palette.
We perform automated surface-level scans (Gen AI Scanner, Socket, Snyk) during installation. These checks detect common vulnerabilities but do not guarantee complete security. Always review skill source code and verify the publisher's reputation before production use.
Skills execute code in your environment. Always review source, verify the publisher, and test in isolation before production.
Submit your Claude Code skill and start earning
Automate repetitive workflows and reduce manual effort
Example
Generate reports, summarize documents, draft communications
Save 3-5 hours per week on routine tasks
Learn new skills, understand complex topics, get expert guidance
Example
Explain concepts, provide examples, suggest learning resources
Accelerate learning and skill development by 2x
Enhance output quality through reviews, suggestions, and refinements
Example
Review drafts, suggest improvements, catch errors
Improve work quality by 30-40% with less effort
0
total installs
0
this week
42
GitHub stars
0
upvotes
Run in your terminal
0
installs
0
this week
42
stars
Manage project documentation for Claude Code workflows including context files, READMEs, and agent instructions. This skill provides structured automation for maintaining accurate, up-to-date documentation that aligns with actual codebase structure and functionality. Use this skill when initializing new projects, updating existing documentation, or ensuring context files accurately reflect current code.
The skill emphasizes verification and validation over blind generation—analyze the actual codebase structure, file contents, and patterns before creating or updating documentation. All generated content should be terse, imperative, and expert-to-expert rather than verbose or tutorial-style.
Before using any documentation workflow, verify basic project structure:
git rev-parse --git-dir
Ensure the output confirms you are in a git repository. If not initialized, documentation workflows may still proceed but git-specific features will be skipped.
For update workflows, verify target files exist:
ls -la CLAUDE.md AGENTS.md DOCS.md README.md CONTRIBUTING.md
Check which files are present before attempting updates. Missing files will show errors, which helps identify what needs initialization. Note that DOCS.md is optional and only relevant for projects with APIs or public interfaces. CONTRIBUTING.md is optional and only relevant if the repo already defines contribution guidance.
Verify and fix AGENTS.md, and optionally DOCS.md and CONTRIBUTING.md against the actual codebase. CLAUDE.md is a symlink to AGENTS.md and does not need separate processing. This workflow reads existing context files, analyzes the codebase structure, identifies discrepancies, and updates documentation to match reality. DOCS.md and CONTRIBUTING.md are only processed if they exist in the repository.
Parse Arguments
Support the following arguments:
--dry-run: Show what would change without writing files--preserve: Keep existing content structure, only fix inaccuracies--thorough: Perform deep analysis of all files (slower but comprehensive)--minimal: Quick verification focusing on high-level structure onlyVerify Git Repository
Confirm working directory is a git repository. If not, warn the user but proceed with limitations (cannot analyze git history or branches).
Read Existing Context Files
Read current AGENTS.md, DOCS.md, and CONTRIBUTING.md (if present) contents. CLAUDE.md is a symlink to AGENTS.md and does not need a separate read:
cat AGENTS.md
cat DOCS.md # if exists
cat CONTRIBUTING.md # if exists
Parse the structure and extract documented information including:
Analyze Codebase
Scan the project to gather accurate information:
ls -la, tree if available)package.json, pyproject.toml, Cargo.toml, etc.)For --thorough mode, also analyze:
Identify Discrepancies
Compare documented information against actual codebase:
CLAUDE.md symlink (should point to AGENTS.md)Create Backups
Before overwriting, create backup files. CLAUDE.md is a symlink to AGENTS.md and does not need a separate backup:
cp AGENTS.md AGENTS.md.backup
test -f DOCS.md && cp DOCS.md DOCS.md.backup
test -f CONTRIBUTING.md && cp CONTRIBUTING.md CONTRIBUTING.md.backup
Update Context Files
Write corrected versions maintaining the existing structure when --preserve is used, or reorganizing for clarity when not. Ensure the CLAUDE.md symlink exists (ln -sf AGENTS.md CLAUDE.md). For --dry-run, display the diff without writing:
diff -u AGENTS.md.backup AGENTS.md
Generate Report
Display a summary of changes.
When DOCS.md exists:
✓ Updated AGENTS.md
- Fixed outdated build command
- Added new /api directory to structure
- Updated test-runner trigger pattern
✓ Updated DOCS.md
- Fixed outdated endpoint path /api/v1/users
- Updated function signature for createUser()
✓ Updated CONTRIBUTING.md
- Updated branch naming convention
- Fixed outdated PR template reference
✓ CLAUDE.md symlink verified
When optional files are absent:
✓ Updated AGENTS.md
- Fixed outdated build command
✓ CLAUDE.md symlink verified
⊘ DOCS.md not found (skipped)
⊘ CONTRIBUTING.md not found (skipped)
For the complete update context files workflow with verification strategies, diff examples, and edge cases, refer to references/update-agents.md.
Generate or update README.md based on project structure, package metadata, and codebase analysis. This workflow creates comprehensive, accurate READMEs that reflect the actual state of the project.
Parse Arguments
Support the following arguments:
--dry-run: Preview README content without writing--preserve: Keep existing sections, only update outdated information--minimal: Generate minimal README (title, description, installation, usage)--full: Generate comprehensive README with all optional sectionsAnalyze Project Structure
Gather information from multiple sources:
# Package metadata
cat package.json
cat pyproject.toml
cat Cargo.toml
# Git information
git remote get-url origin
git describe --tags
# Directory structure
ls -la
Extract:
Read Existing README
If README.md exists and --preserve is used:
cat README.md
Parse existing sections to preserve custom content while updating technical details.
Read CONTRIBUTING.md (Optional)
If CONTRIBUTING.md exists, read it and treat it as the source of truth for contribution guidance. Do not duplicate detailed contribution steps in README; link to CONTRIBUTING.md instead.
test -f CONTRIBUTING.md && cat CONTRIBUTING.md
Create Backup
Before overwriting existing README:
cp README.md README.md.backup
Generate README Content
Create structured content with appropriate sections:
For --minimal mode, include only title, description, installation, and usage.
For --full mode, also include:
Write README
Save the generated content. For --dry-run, display without writing.
Generate Report
Display summary:
✓ Updated README.md
- Added installation section
- Updated build commands to match package.json
- Added badges for license and version
For the complete update README workflow with section templates, metadata extraction strategies, and formatting examples, refer to references/update-readme.md.
Update CONTRIBUTING.md based on current codebase tooling and workflows. This workflow only runs when CONTRIBUTING.md already exists in the repository. If CONTRIBUTING.md is absent, skip this workflow entirely—do not auto-create contribution guidelines.
Before proceeding, verify the file exists:
test -f CONTRIBUTING.md && echo "exists" || echo "missing"
If missing, report to the user and stop. Do not create CONTRIBUTING.md unless explicitly requested.
Parse Arguments
Support the following arguments:
--dry-run: Show what would change without writing files--preserve: Maximum preservation; only fix broken commands/links--thorough: Deep analysis; verify all links and commands workRead Existing CONTRIBUTING.md
cat CONTRIBUTING.md
Parse the document structure:
Gather Codebase Intelligence
Detect current tooling and compare against documented content:
Identify Discrepancies
Compare documented information against actual codebase:
Update Content
Fix technical inaccuracies while preserving:
Generate Report
Display summary:
✓ Updated CONTRIBUTING.md
- Fixed package manager: npm → pnpm
- Corrected branch reference: master → main
- Updated test command
⊘ Policy sections preserved (CLA, review process)
For the complete update CONTRIBUTING workflow with verification strategies and examples, refer to references/update-contributing.md.
Create project-specific AGENTS.md from scratch based on codebase analysis. This workflow is ideal for new projects or repositories lacking context documentation.
Parse Arguments
Support the following arguments:
--dry-run: Preview generated content without writing--minimal: Create minimal context file (project description, structure)--full: Create comprehensive context file with all relevant sectionsVerify No Existing AGENTS.md
Check if AGENTS.md already exists:
test -f AGENTS.md && echo "exists" || echo "missing"
If exists, warn the user and suggest using the update workflow instead. Allow override with --force flag.
Analyze Project
Gather comprehensive information:
Generate AGENTS.md Content
Create structured sections:
# Context
Brief project description and purpose.
## Structure
Directory organization and key files.
## Build
Commands for building the project.
## Test
Commands for running tests.
## Development
Conventions, patterns, and workflows.
Adapt sections based on project type. For --minimal, include only Context and Structure. For --full, add all applicable sections including deployment, troubleshooting, and custom tooling.
Write AGENTS.md
Save generated content and create CLAUDE.md symlink (ln -sf AGENTS.md CLAUDE.md). For --dry-run, display without writing.
Generate Report
Display summary:
✓ Created AGENTS.md
- Detected Next.js project
- Added npm scripts from package.json
- Documented project structure
- Added testing section for Jest
✓ Created CLAUDE.md symlink
For the complete initialize context workflow with language-specific templates, detection strategies, and customization options, refer to references/init-agents.md.
DOCS.md is optional and not created by default. Create DOCS.md manually when the project has:
The update context workflow will suggest creating DOCS.md if it detects significant APIs without corresponding documentation.
For detailed workflows, examples, and implementation guidance, refer to these reference documents:
references/common-patterns.md - Argument parsing, backup handling, writing style, report formatting, file detection, and metadata extractionreferences/update-agents.md - Complete context file update workflow including verification strategies, diff generation, and discrepancy detectionreferences/update-readme.md - Complete README update workflow including section templates, metadata extraction, and formatting conventionsreferences/update-contributing.md - Complete CONTRIBUTING.md update workflow including scope, templates, and validation (only when CONTRIBUTING.md exists)references/init-agents.md - Complete context initialization workflow including language-specific templates, detection strategies, and customization optionsThese references provide implementation details, code examples, and troubleshooting guidance for each workflow type.
Prerequisites
Time Estimate
15-45 minutes depending on use case complexity
Steps
Common Pitfalls
✓ Do
✗ Don't
💡 Pro Tips
✓ Use when
Use when skill capabilities match your task, clear ROI on time saved, and you can validate outputs. Best for repetitive tasks, learning, and quality improvement.
✗ Avoid when
Avoid when task requires deep expertise you can't validate, involves sensitive decisions, or when learning process is more valuable than speed of completion.
langchain-ai/deepagents
upstash/context7
google-gemini/gemini-cli
zrong/skills
lombiq/tailwind-agent-skills
patricio0312rev/skills
Solid pick for teams standardizing on skills: md-docs is focused, and the summary matches what you get after install.
We added md-docs from the explainx registry; install was straightforward and the SKILL.md answered most questions upfront.
md-docs is among the better-maintained entries we tried; worth keeping pinned for repeat workflows.
Useful defaults in md-docs — fewer surprises than typical one-off scripts, and it plays nicely with `npx skills` flows.
I recommend md-docs for anyone iterating fast on agent tooling; clear intent and a small, reviewable surface area.
md-docs is among the better-maintained entries we tried; worth keeping pinned for repeat workflows.
md-docs fits our agent workflows well — practical, well scoped, and easy to wire into existing repos.
Keeps context tight: md-docs is the kind of skill you can hand to a new teammate without a long onboarding doc.
Solid pick for teams standardizing on skills: md-docs is focused, and the summary matches what you get after install.
md-docs is among the better-maintained entries we tried; worth keeping pinned for repeat workflows.
showing 1-10 of 41