Skip to content

CONTRIBUTING.md

Latest commit

 

History

History
40 lines (28 loc) · 1.34 KB

CONTRIBUTING.md

File metadata and controls

40 lines (28 loc) · 1.34 KB
Raw

Contributing to Helm

Helm is an operations layer for long-lived AI agent workspaces. Contributions are most useful when they improve safety, recovery, auditability, or durable context for repeated agent work.

Good First Contributions

  • Improve first-run docs, examples, or comparison pages.
  • Add tests around guardrail behavior, checkpoint handling, reports, or workspace detection.
  • Make CLI output clearer without changing machine-readable formats.
  • Add portable integrations for existing local agent workspaces.

Design Principles

  • Keep Helm runtime-agnostic. It should work around agent systems rather than becoming one.
  • Prefer files as the source of truth. SQLite and reports are indexes or views.
  • Fail closed for risky command handling and policy parsing.
  • Keep private memory, credentials, and personal overlays out of the public project.

Development

python3 -m pytest -q
git diff --check

For packaging changes, also run a target install smoke test:

python3 -m pip install --target /tmp/helm-install-check .
env PYTHONPATH=/tmp/helm-install-check /tmp/helm-install-check/bin/helm init --path /tmp/helm-installed-test --force

Pull Requests

Please include:

  • the user-facing problem being solved
  • the affected commands or docs
  • tests or smoke checks run
  • any compatibility notes for existing workspaces