delegate-investigation-tasks-under-command
Use when the incident commander must task responders without joining the debugging. Assigns bounded, time-boxed investigation threads with a defined return format so command stays unblocked.
- License 许可
- MIT
- 最近更新
- 2026-11-12
- Stars
- ★ 81
- Forks
- 17
01 前言 Front Matter
- name
- delegate-investigation-tasks-under-command
- description
- Use when the incident commander must task responders without joining the debugging. Assigns bounded, time-boxed investigation threads with a defined return format so command stays unblocked.
02 资源文件 3
- 许可证LICENSE1.0 KB
- 参考文档SOURCE.md368 B
- 参考文档TRUST.auto.yaml2.2 KB
03 SKILL.md 正文
Delegate Investigation Tasks Under Command
An IC who starts grepping logs has stopped commanding. Delegation keeps command on the decision layer while responders work bounded threads — but only if each task has an owner, a timebox, and a defined form of answer.
Procedure
- Frame each task as a question with a test, not an activity: not "look at the DB" but "in 10 minutes tell me whether active DB connections stay high after traffic drops — that distinguishes leak from load."
- Give every task an owner and a timebox: "@ben, 10 minutes, report by 14:25."
- Specify the return format so the answer is usable: "one line: connections after the drop, and whether that supports the leak."
- Keep the IC out of the thread; if the IC must investigate, they hand command to a deputy first.
- Limit parallelism: 3-4 open threads max; more and the IC cannot hold the state and handoffs degrade.
- Close each thread explicitly when answered — "leak confirmed, thanks, drop it" — so responders are not left digging in a dead end.
- Track threads somewhere visible (the channel) so a handoff at shift change does not strand one.
- Write the outstanding threads into the handoff block when the IC changes, or they are lost.
- Give the responder the evidence they need to start (log path, dashboard link) rather than making them hunt.
Pitfalls
- Assigning an activity ("investigate the cache") with no question, so the responder returns raw data and no conclusion.
- Timeboxes that are never checked, so a 10-minute task runs an hour and blocks the branch that depended on it.
- The IC picking up a keyboard "just to check", silently abandoning command.
- Six parallel threads the IC cannot reconcile, so the same hypothesis is tested twice and two others are dropped.
- Never closing threads, so the responder keeps working on a question the team already answered.
- Delegating without the return format, so five people answer in five shapes the IC must translate.
Verification
grep -E '@\w+, [0-9]+ min, report by' incident/SEV*-2026-*.md
# passes when every delegated thread names an owner, a timebox, and a specific return question
Related: run-an-incident-call-bridge covers how the IC voices these tasks without cross-talk.
04 同仓库更多 Skills
recipe-performance
Diagnose a specific MetaMask user flow's real-app performance and Sentry critical-user-flow telemetry with deterministic mm-harness recipes. Use for CUF boundaries, cold/warm/background-resume behavior, latency regressions, lifecycle tags, duplicate or blocking network requests, long tasks, layout/paint cost, screen redraws, React rerenders, or browser trace analysis where tests alone cannot prove the real app.
mexc-spot-websocket
MEXC Spot WebSocket workflows for public streams, protobuf handling, depth maintenance, listen keys, and private streams without exposing API keys.
mexc-spot-rest
MEXC Spot REST workflows for public data, signed account requests, orders, wallet/admin actions, and secure signing without exposing API keys.
mexc-futures-websocket
MEXC Futures WebSocket workflows for public streams, private account/order/position streams, ping/pong, subscriptions, and depth maintenance.
0xklen-version-and-refresh-sanctions-lists
暂无简介
verify-tests-cover-changed-lines
Use when a PR adds code that the existing suite may not exercise. Measures patch coverage — the share of newly changed lines hit by tests — before approving.