minion-orchestrator

Orchestrate durable background shell jobs and LLM subagent runs with lifecycle controls.

5|1|Updated Apr 15, 2026
One-click install
npx skills add https://github.com/GYF0311/lorekit --skill minion-orchestrator-gyf0311
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: minion-orchestrator
Source: https://github.com/GYF0311/lorekit/tree/main/brain/skills/minion-orchestrator
Command: npx skills add https://github.com/GYF0311/lorekit --skill minion-orchestrator-gyf0311

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It eliminates brittle, non-durable background execution and ad-hoc orchestration by providing a single unified queue interface for both deterministic shell work and LLM subagent runs.

Core Features & Use Cases

  • Deterministic shell job routing: Schedule and monitor Postgres-backed background command execution with structured status, progress, and transcripts.
  • LLM subagent orchestration: Submit open-ended agent runs (including fan-out with an aggregator), then monitor tokens, steps, and outcomes via job details and progress snapshots.
  • Full lifecycle control: Steer, pause, resume, cancel, and replay jobs while preserving durability and observability guarantees.

Quick Start

Ask your AI to submit a durable background agent run for a research prompt: "Use minion-orchestrator to run 'Research Acme Corp revenue' as a Minion subagent job and stream progress."

Frequently Asked Questions about minion-orchestrator

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run durable background execution for shell commands and LLM subagents?

Durable background execution is handled by submitting deterministic shell jobs and LLM subagent workloads to a unified queue interface, ensuring work survives restarts and remains observable through structured status and progress reporting.

What is job orchestration for fan-out parallelism and how does it work?

Job orchestration for fan-out parallelism allows submitting open-ended subagent runs that execute concurrently, with an aggregator collecting outcomes while monitoring tokens, steps, and results via detailed progress snapshots.

Can I pause, resume, cancel, or replay background jobs after submission?

Full lifecycle control allows you to steer, pause, resume, cancel, and replay background jobs while preserving durability and observability guarantees across both deterministic shell commands and LLM subagent runs.

Does minion-orchestrator work with Postgres-backed queue monitoring for background tasks?

Postgres-backed queue monitoring is supported for deterministic shell job routing, providing structured status, progress, and transcripts for scheduled background command execution.

What are the limitations of using allow-listed tools for LLM subagent orchestration?

LLM subagent orchestration requires routing rules that enforce shell-job submission restrictions and allow-listed tool usage, meaning subagent runs are constrained to explicitly permitted tools for safety and deterministic execution.

How do I monitor tokens and steps for LLM subagent jobs?

Monitoring tokens and steps for LLM subagent jobs is done through job details and progress snapshots exposed via MCP-callable queue operations, providing structured reporting on open-ended agent run outcomes.