- Home
- Skills
- Prompt & Skill Engineering
- project.init
project.init
Drafts PRD.md (product requirements), TRD.md (technical design), and EXECUTED.md (append-only execution log) at the project root 3.
Free
Works with the AI Agents you already use
project.init
Example session with this skill installed
/project-init
then answers its 4 questions, e.g.:
name: acme-api · purpose: "REST API for order tracking"
mode: full · author: Jane Doe · extras: README, MIT LICENSE, initial commit
- Read your context and instructions
- Compiled the project init
A committed git repo scaffold
acme-api/
├── PRD.md # problem, goals, personas, P0–P2 requirements, metrics
├── TRD.md # architecture, stack, data model, APIs, security, milestones
├── EXECUTED.md # execution log — first entry = this init (ISO ts + agent/model)
├── AGENTS.md # conventions every future agent must follow
├── CLAUDE.md / GEMINI.md # pointers → AGENTS.md
├── README.md .gitignore LICENSE
└── docs/
├── README.md # KB index
├── ai/ technical/ layman/ non-technical/ ← 4 seeded audience tracks
Every folder gets a README.md; every source file gets:
/**
- Handles OAuth token refresh.
- Docs: docs/technical/api.md#auth
- Author: Jane Doe
- Last-Updated: 2026-11-14T10:35:00+05:30
*/
Connects securely to your tools. The creator never sees your data.
About this skill
project-init bootstraps any software project to a universal documentation standard in one command.
What it does
- Runs git init (branch: main) and writes a baseline .gitignore
- Drafts PRD.md (product requirements), TRD.md (technical design), and EXECUTED.md (append-only execution log) at the project root
- Builds a docs/ knowledge base in four audience tracks — ai/ (machine-optimized), technical/ (engineers), layman/ (plain language), non-technical/ (business framing) — with an index README
- Drops an AGENTS.md conventions file (plus CLAUDE.md / GEMINI.md pointers) so every future AI session keeps docs in sync automatically
- Full mode: writes a README.md for every folder and prepends a header comment to every source file — one-line purpose, a Docs: link into docs/, Author, and an ISO-8601 Last-Updated timestamp
Modes
- init (default) — full setup for a new project
- upgrade / full — idempotent retrofit of an existing repo; fills gaps, never overwrites
- lite — skeleton docs now; re-invoke upgrade later
It always asks first: project name, one-line purpose, mode, and author (defaults to git config user.name). Nothing is fabricated — unknowns become TODO markers.
Built on the open SKILL.md standard, so one package works everywhere: Claude Code, Codex, Cursor, VS Code Copilot, Devin, Antigravity, Gemini CLI, Windsurf, Kiro, Amp, OpenCode, Zed, Cline, Roo and more. No API keys, no telemetry, MIT licensed.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Prompt & Skill Engineering.
• Auto-invocation on "new project" phrasing depends on the host agent; manual /project-init always works.
• Files that can't hold comments (JSON, lockfiles, binaries, generated code) get no header — they're covered by the folder README instead.
• full mode on very large repos proceeds folder-by-folder and asks the user to scope.
• Generated doc quality depends on the underlying model; anything it can't infer becomes a TODO marker.
• Vendored / third-party / generated code is intentionally out of header scope.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 3 days ago
- Free with an Agensi account
- Software
Gitshell
Python
Permissions
The skill always asks 4 setup questions before writing anything. Full mode touches every source file to add headers — run it on a fresh repo or a clean branch. It never deletes content (merge/append only) and is fully idempotent — safe to re-run.