Skip to content
Back to skills

Project Docs

ASecurity

Audit a project for the agent doc set (AGENTS.md, CLAUDE.md, ARCHITECTURE, README, CODE_STYLE, TESTING, SECURITY, PRD, DESIGN_SYSTEM), report gaps, draft missing ones from the code.

  • 5 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added November 11, 2026
ai-agentsrustgotestinggitfrontendsecurity

Works with

  • cursor
  • cli

Security analysis

A100/100

Scanned November 11, 2026

npx -y skills add ahmadhammoud56/MOAA --skill project-docs --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Project Docs?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Project Docs
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/ahmadhammoud56-project-docs/badge)](https://www.skillsdirectory.com/skills/ahmadhammoud56-project-docs)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: project-docs
description: Audit a project for the agent doc set (AGENTS.md, CLAUDE.md, ARCHITECTURE, README, CODE_STYLE, TESTING, SECURITY, PRD, DESIGN_SYSTEM), report gaps, draft missing ones from the code.
---

# Project docs

Request: `$ARGUMENTS` (empty: audit the current project).

## 1. Audit

Look in the project root (and `docs/`) for each file, case-insensitive. Mark each **present**, **missing**, or
**stale** (names files, commands or folders that no longer exist, or its last commit is far behind the code it
describes: `git log -1 --format=%cr -- <file>` against `git log -1 --format=%cr -- <src dir>`).

| File | Purpose | Expect it when |
|---|---|---|
| `AGENTS.md` | Single source of agent rules: purpose, before you start, rules, code guidelines, security, useful commands | always |
| `CLAUDE.md` | One line, `@AGENTS.md`, plus Claude-only notes | always |
| `README.md` | What it is, install, quick start | always |
| `ARCHITECTURE.md` | Map of components, data flow, entry points | more than a handful of modules |
| `CODE_STYLE.md` | Conventions the linter does not enforce | team or multi-agent work |
| `TESTING.md` | How to run tests, where they live, what must pass | there are tests |
| `SECURITY.md` | Secrets handling, trust boundaries, reporting | network, auth, user data, or public repo |
| `PRD.md` | Goals, users, features, success metrics | app or product only |
| `DESIGN_SYSTEM.md` | Colours, type, spacing, components | UI project only |

Suggest `PRD.md` and `DESIGN_SYSTEM.md` only for app or UI projects (a frontend framework, `src/components`,
CSS/Tailwind, mobile app). Short sections of CODE_STYLE or TESTING can live inside AGENTS.md instead of their own file.

Report one table (file, status, one-line reason) and stop. Ask which missing or stale files to draft.

## 2. Draft (only the files the user picks)

- Write from the code itself: real commands from `package.json`, `pyproject.toml`, `Makefile`, CI files; real folder
  names; real test runners. Never invent features, metrics or policies; leave a `TODO:` where only the user knows.
- **AGENTS.md is the single cross-agent source.** Codex, Gemini CLI, OpenCode and Cursor read it. Keep `CLAUDE.md`
  to `@AGENTS.md` plus anything only Claude needs. If a long CLAUDE.md exists, offer to move its shared rules into
  AGENTS.md and leave the import line; do not duplicate text between the two.
- Keep each file short (AGENTS.md under about 150 lines); link to the others rather than repeating them.

Use an installed skill where one fits instead of writing from scratch:

| File | Skill |
|---|---|
| `AGENTS.md` | `agentic-awesome-skills:agents-md` |
| `PRD.md` | `ecc:plan-prd` |
| `DESIGN_SYSTEM.md` | `agentic-awesome-skills:design-md` |
| `ARCHITECTURE.md`, codemaps | `ecc:update-codemaps` |

If a skill is not installed, draft the file directly following the table in step 1. Show each draft as a diff or new
file and let the user approve before writing over anything that exists.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

View all in ai-agents →

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.