Agent Dispatch

User-invoked only: run it when the user explicitly asks to hand work to an external agent CLI, never on your own initiative. On enable it probes the machine first — which agent CLIs are on PATH, and which LLM each one runs, read from that CLI's own --help / --list-models, never from memory or this file — then asks the user, one question at a time, to pick the agents, their model modes, and whether to append. It drives the loop after that: a task brief carrying only checked-in artifact paths plus an output format, an isolated worktree, a background run, and admission only after the calling project's own gates pass. Four rules hold whoever is picked: no self-review, outputs gated before merge, briefs carry no secrets or handoff stores, and the current driver may do work but never audits this session's output. Use for 派给外部 agent / 换个模型来审 / 调遣 agent. NOT for: doing the work yourself, or accepting the output (that stays with the caller), or dispatching where no gates were declared — ask first.

NinjaSln-labs Updated 4 repo stars

File contents

ninjasln-labs/agent-skills commit 8dfcc3e0b1

Frequently asked questions

npx get-skill add ninjasln-labs/agent-dispatch