Skip to content

Latest commit

 

History

History
49 lines (37 loc) · 1.68 KB

File metadata and controls

49 lines (37 loc) · 1.68 KB

Contributing

Produced by Alejandro Pinto.

Contact: super.mangmail@gmail.com

IRIS is currently open for bug fixes, diagnostics fixes, documentation fixes, compatibility fixes, and safety-preserving test coverage.

Allowed Contributions

  • Bug fixes.
  • Failing-test fixes.
  • Local Windows compatibility fixes.
  • Documentation corrections.
  • Security hardening that preserves Safe mode and Agentic approval boundaries.
  • Test coverage for existing behavior.

Avoid Without Explicit Approval

  • New action tools outside the reviewed Agentic profile.
  • Mouse, keyboard, clipboard, accessibility, or general window control.
  • Browser, shell, or process capabilities that bypass Agentic session approval, per-action high-risk confirmation, audit logging, or Panic Stop.
  • Model switching, model pulling, fallback models, or multi-model debate.
  • External/cloud network behavior.
  • New dependencies.
  • UI redesigns.
  • Feature expansions beyond the current safety envelope.
  • Changes that weaken Hermes restrictions.
  • Live memory storage under cloud-sync folders.

Validation Before a PR

Run from C:\Projects\IRIS:

cargo fmt --all -- --check
cargo build --workspace
cargo test --workspace
cargo clippy --workspace
cargo run -p xtask
npm run test:voice
npm run test:python
git diff --check

Manual runtime testing is described in docs/manual-test.md.

GitHub Actions runs the lightweight bug checker on pushes and pull requests. That CI validates source, tests, and repository audits, but it does not launch the desktop runtime, use a microphone, speak audio, access a camera, capture the screen, or call Ollama. Manual runtime testing still matters for changes that affect the app loop.