tackroom docs

Guides and reference for the tackroom CLI, in reading order. Philosophy says what tackroom is for; setup gets you running; the rest stand alone.

Start here

  • Philosophy: what tackroom believes and refuses to do.
  • Setup: first run, --review, more machines with --from and --overlay, settings, root instructions, upgrades with the release notes, and uninstalling.
  • Commands: every command by task, exit codes, config lookup order and output rules.
  • Harnesses: what each of the 16 agents gets, the key that names it and where its files land.

Guides

  • Skills: write your own, add external sources, provenance, the 72 hour cooldown, the audit gate and publishing.
  • Roles: the role file format, where each harness gets it, model tiers and overrides.
  • MCP servers: add, import and remove servers, remote servers, secrets and which harnesses get what.
  • Hooks: declare a hook, the events each harness takes, and list or remove native registrations.
  • Plugins: Claude Code plugins, agent-plugins-spec sources and the Claude Code and Pi panels.
  • Memory: deliver a memory file to every harness, native memory settings, and separate tools such as rem.
  • Crons: inventory every job on every host, run and pause launchd jobs, declare your own.
  • Web view: the loopback dashboard, its token, remote access and what it can change.
  • Terminals: tackroom dash, the tmux, zellij, Tern and Rex workspaces that open it, the Herdr panel and a cmux recipe.

More

  • Comparison: how tackroom differs from skillshare, HarnessKit, dotagents, rulesync, ruler, openskills, npx skills and APM, and how to move from skillshare.
  • Troubleshooting: conflicts, items in cooldown and blocked items, hooks, Pi, periodic sync.

Edit this page on GitHub