Pynchy Development

Pynchy Development

Run commands directly—don't tell the user to run them.

Am I on pynchy?

Check hostname and compare it with the deployment-specific live host from local memory, environment, or the operator. If you are on that host, access services at localhost. Otherwise, reach Pynchy over SSH or Tailscale using that configured host.

Commands

uv run pynchy            # Run the app
uv run pytest tests/     # Run tests
uv run ruff check --fix src/  # Lint + autofix
uv run ruff format src/       # Format
uvx prek==0.4.10 run --all-files  # Run all repository hooks
./src/pynchy/agent/build.sh     # Rebuild agent container

Managed worktrees and optional runtimes

Use new-feature create <slug> --no-agent from the control checkout before any source or documentation edit. Work in the resulting .worktrees/<slug> directory; direct edits on the target branch are blocked. Setup installs dependencies but does not start a runtime. Start an isolated Pynchy runtime only when interactive diagnosis needs one. See the feature-runtime workflow for create, merge, restart, and teardown commands. Do not use raw git worktree commands for managed features. Install or verify its host dependencies with ./scripts/install_new_feature_dependencies.py.

Private overlay edits belong in the independent personalization repository's own managed worktree, not in the public Pynchy worktree. See editing personalization.

Quality checks

Keep correctness, security, architecture, and coverage checks blocking. The comment-language check is advisory. Host beartype violations warn in production and fail tests; fix the contract or test double instead of suppressing violations. Host annotations used by beartype need runtime imports; Ruff's host configuration handles this without per-import exemptions. Agent-runner uses its own lint config.

Run focused checks while editing a coherent change. The managed merge owns the full verification sequence; repeat it only after changes or unresolved failures.