Skip to content
Get packs

Minion Orchestrator

Runs and tracks durable background jobs, shell commands and LLM subagents, with steering and deadman checks.

Source Download skill
Repository
garrytan/gbrain
Stars
31,447
Licence
MIT
Files
1 file

What it does

This skill teaches an AI agent how to hand work to a background job queue (called Minions, backed by Postgres) instead of running it inline. Long or parallel tasks can then survive restarts, be checked on, paused, resumed, or redirected, and the agent follows rules to confirm the result actually reached the user. Shell command jobs can only be submitted by an operator via the CLI; the agent can observe them but not submit them over MCP.

Great for

  • Developers running the gbrain tool who need long ETL, cron, or pipeline steps to run durably in the background
  • Operators who manage a shared job queue and want agents to monitor and steer subagent work
  • Anyone running multi-step research or fan-out tasks that must not silently die with the session

Skip it if

  • Quick one-off commands under about 30 seconds, where inline execution is simpler and safer; the skill itself suggests this
  • Shared or multi-user queues where enabling shell jobs would let arbitrary commands run on the worker, a remote code execution risk, unless access is tightly controlled
  • Users without the gbrain project and Postgres or PGLite set up, since the skill only works with that tooling

Try asking

  • “Submit a shell job to run my nightly ETL script in the background and tell me when it finishes”
  • “What's running right now? Check on the subagent I started this morning”
  • “Fan out these five research questions in parallel and report back when all are done”
  • “Pause the long data pipeline and resume it after the database migration”
Step-by-step guide: how to use Minion OrchestratorInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the MIT licence and an attribution note.

    minion-orchestrator.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip minion-orchestrator.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip minion-orchestrator.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Enable shell jobs only on a private worker you control, and treat GBRAIN_ALLOW_SHELL_JOBS as privileged infrastructure access.
  • Run gbrain jobs stats after setup to confirm a worker is registered; jobs sitting in waiting mean no worker is running.
  • Ask the agent to set a deadman check for long operations so you are notified if a result is never reported.

What's inside

  • SKILL.md

Redistributed unmodified from garrytan/gbrain under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.