Agent Instructions Kit

Version updated for https://github.com/malmichaels-gif/agent-instructions-kit to version v0.1.0.

• This action is used across all versions by ? repositories.

Go to the GitHub Marketplace to find the latest changes.

Action Summary

The agent-instructions-kit GitHub Action automates the creation, validation, and safety checks of agent instruction files (e.g., AGENTS.md and CLAUDE.md) used by AI coding tools. It helps maintain consistency, enforces required sections, and detects potential security issues such as prompt injection risks or unsafe patterns. This action streamlines the management of agent instructions and ensures they remain secure, up-to-date, and well-structured.

Release notes

What’s New

Initial release of agent-instructions-kit.

Features

• init: Generate AGENTS.md + CLAUDE.md from templates (minimal or opinionated)
• check: Validate required sections exist
• safety: Lint for suspicious patterns (prompt injection, secret exfiltration)
• GitHub Action: Run checks in CI with configurable fail behavior

CLI Usage

npx agent-instructions-kit init
npx agent-instructions-kit check
npx agent-instructions-kit safety

Action Usage

- uses: malmichaels-gif/agent-instructions-kit@v0
  with:
    mode: "all"
    fail_on_safety: "false"

See README for full documentation.

• malmichaels-gif
• GitHub Actions