init

Create an AGENTS.md contributor guide for a repository (or a subdirectory) following a concise, repo-specific template. Use when the user runs /init.

Safety Notice

This listing is imported from skills.sh public index metadata. Review upstream SKILL.md and repository scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "init" with this command: npx skills add openhands/skills/openhands-skills-init

Initialize AGENTS.md

Use this skill when the user asks to scaffold an AGENTS.md file (project guidelines for AI agents) for the current repository.

Target path rules

  • /init: create AGENTS.md at the repository root.
  • /init <path>: create AGENTS.md scoped to <path>.
    • If <path> is a directory, write <path>/AGENTS.md.
    • If <path> ends with .md, treat it as the full output file path.

Never overwrite an existing file silently. If the target file already exists, ask the user whether to (a) keep it, (b) edit it, or (c) replace it.

Workflow

  1. Inspect the repo to tailor the guide

    • Skim top-level structure (ls, find -maxdepth 2).
    • Read the primary docs (README*, CONTRIBUTING*, DEVELOPMENT*, Makefile).
    • Identify build tooling and common commands (pyproject.toml, package.json, CI configs).
    • Locate tests and how they are run (tests/, workflow YAMLs, task runners).
  2. Write a concise AGENTS.md using the spec below.

  3. Create the file at the target location.

AGENTS.md spec (inspired by Codex /init)

  • Title the document: # Repository Guidelines.
  • Use Markdown headings (##) to organize sections.
  • Keep it concise: 200–400 words.
  • Keep explanations short, direct, and specific to the repo.
  • Prefer concrete examples (commands, directory paths, naming patterns).

Recommended sections (adapt as needed)

  • Project Structure & Module Organization
  • Build, Test, and Development Commands
  • Coding Style & Naming Conventions
  • Testing Guidelines
  • Commit & Pull Request Guidelines
  • Optional: Security & Configuration Tips, Architecture Overview, Agent-specific instructions

Content guidelines

  • Only include information that’s broadly useful for future tasks (avoid one-off issue details).
  • If the repo already contains nested AGENTS.md files, briefly mention that more deeply nested files override broader ones within their directory tree.

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

Automation

agent-memory

No summary provided by upstream source.

Repository SourceNeeds Review
Automation

onboarding-agent

No summary provided by upstream source.

Repository SourceNeeds Review
General

ssh

No summary provided by upstream source.

Repository SourceNeeds Review
Coding

codereview-roasted

No summary provided by upstream source.

Repository SourceNeeds Review