Thank you for your interest in contributing! Engram is a lightweight protocol — the best contributions are ones that keep it that way.
Open an issue using the Bug Report template. Include:
- What you expected to happen
- What actually happened
- The AI model and interface you were using (Claude, ChatGPT, Cursor, etc.)
Open an issue using the Feature Request template. Explain the use case first — before proposing the implementation.
If you've adapted Engram for a specific workflow (academic research, architecture, creative writing, autonomous agents), open an issue describing it. Validated workflows are incorporated into the example library.
- Fork the repo
- Create a feature branch:
git checkout -b feature/my-improvement - Make your changes
- Test: run
./init-engram.shin a temp directory and verify the output - Submit a Pull Request against
mainwith a clear description of what changed and why
✅ Smaller, focused PRs over large rewrites
✅ Improvements that work across all AI platforms (Claude, ChatGPT, Gemini, Cursor)
✅ Additions to examples/ showing real project use
✅ Fixes to workflow files that improve reliability or clarity
❌ Changes that add external dependencies
❌ Rewrites of the logging format without a migration path
❌ Platform-specific features that don't degrade gracefully on other platforms
The .agents/workflows/ directory is the most impactful place to contribute. Each workflow is a plain markdown file with a YAML frontmatter description. Changes there affect every AI model that uses Engram.
Be direct, be honest, and be kind. Flag problems clearly. Give credit where it's due.
This project follows the Contributor Covenant v2.1.
Open an issue or reach out at jim@ecom-x.com.