AGENT SKILL

qhealth

Compact service health check across the portfolio — what's down and what needs action, nothing else. Use before dispatching agents, after a deploy or restart, or when asked whether services are up.

License 许可
MIT
所属仓库
hollis-labs/cairn
最近更新
2026-10-21
Stars
★ 0
Forks
0

01 前言 Front Matter

name
qhealth
description
Compact service health check across the portfolio — what's down and what needs action, nothing else. Use before dispatching agents, after a deploy or restart, or when asked whether services are up.

03 SKILL.md 正文

Run this in a subagent. cerberus_health returns ~35 resources (~4KB) and mux_health adds more. That raw payload must not land in the calling context — the entire point of this skill is a small answer.

Subagent prompt

You are a health reporter. Call mcp__mux__cerberus_health and mcp__mux__mux_health. Return ONLY the block below — no raw output, no commentary.

Report exceptions, not inventory. A list of 30 healthy services is noise.

=== QHEALTH ===
daemon    <up|DOWN>   mux <version>
resources <N> running · <M> stopped · <K> unhealthy

NEEDS ACTION
  <resource_id>  <status>  <recommended_action> — <recommended_reason>
  ...or "none"

STOPPED
  <resource_id> ...   (omit any with operator_stopped: true)
  ...or "none"
===============

A resource with status: running but healthy: false belongs under NEEDS ACTION — usually artifact_stale, meaning the running binary is behind the built artifact. Include its recommended_next_step verbatim if present.

Skip operator_stopped: true resources entirely — those are stopped on purpose.

Output

Print the subagent's block. Add nothing.

Invariants

  • Always via subagent. Never pull the raw health payload into the main context.
  • Read-only. Never apply, restart, or reload anything — report the recommended action, let the user decide.
  • Exceptions only. If everything is healthy, that's three lines.

04 同仓库更多 Skills