# Loom

An AI development orchestration tool that spawns coding agents to claim GitHub or Gitea issues, implement features, review pull requests, and merge code, coordinated entirely through issue and PR labels.

Loom spawns AI agents that claim issues, implement features, review pull requests, and merge code, using labels on a forge (GitHub or Gitea, auto-detected from the git remote) as the coordination layer; the developer's job is reduced to writing issues, reviewing pull requests, and merging what looks good.

Installing it into a repository and running `/loom:sweep <issue>` from Claude Code starts autonomous work on one or more issues in parallel waves, or `/loom:sweep all` works the whole open backlog; a `loom-daemon` handles continuous, multi-account, token-rotated batches for ongoing use. The full workflow is driven by labels such as `loom:issue`, `loom:building`, `loom:review-requested`, and `loom:pr`, with a Rust daemon implementing the state machine for its epic-supervisor lane specifically, including fork-join barriers for multi-issue epics.

Installation is a shell script run against the target repository.

Website: https://github.com/rjwalters/loom