A terminal-based tool for orchestrating multiple parallel AI coding agents, using git worktrees to isolate each agent's work.
Built with Rust and ratatui.
Before starting work, read these docs:
- docs/DESIGN.md - Architecture, data model, and UI layout
- docs/IMPLEMENTATION_PLAN_V3.md - Phased implementation with success criteria
If continuing from a previous phase, also read the relevant handoff doc in .context/ (e.g. .context/handoff-phase-15.md).
- Read the implementation plan and the previous phase's handoff doc.
- Implement the phase, writing tests for all new functionality.
- Verify your work passes all checks before considering the phase complete:
direnv exec . cargo fmt # Format code direnv exec . cargo clippy # Lint (must be warning-free) direnv exec . cargo test # All tests must pass
- Write a handoff doc at
.context/handoff-phase-<N>.mdsummarizing:- What was implemented
- Files changed
- Test summary (count, any notable coverage)
- Any decisions or gotchas for the next phase
- Commit and push your changes.
Commands run through direnv to ensure correct environment:
direnv exec . cargo build # Build
direnv exec . cargo run # Run
direnv exec . cargo test # Test
direnv exec . cargo clippy # Lint
direnv exec . cargo fmt # FormatSee docs/IMPLEMENTATION_PLAN_V3.md for current phase and progress.