bound-an-open-ended-investigation
Use when a task is research-shaped, such as "look into" or "figure out why". Set a time box, an output, and a stop condition up front so investigation cannot absorb the deadline.
- License 许可
- MIT
- 最近更新
- 2026-11-14
- Stars
- ★ 81
- Forks
- 17
01 前言 Front Matter
- name
- bound-an-open-ended-investigation
- description
- Use when a task is research-shaped, such as "look into" or "figure out why". Set a time box, an output, and a stop condition up front so investigation cannot absorb the deadline.
02 资源文件 3
- 许可证LICENSE1.0 KB
- 参考文档SOURCE.md359 B
- 参考文档TRUST.auto.yaml2.2 KB
03 SKILL.md 正文
Bound an open-ended investigation
"Look into it" has no natural end, so it expands to fill whatever time exists. A time box, a fixed output, and a stop condition give it a boundary that forces a decision.
Procedure
-
Rewrite the ask as a question with a decision attached: "why is p99 slow" becomes "decide whether to rewrite the cache".
-
Set a time box before starting: 90 minutes, or 10% of the remaining budget.
-
Fix the output shape up front: a one-page finding, a recommendation, or a Go/No-Go. Not "more investigation".
-
Define the stop condition: the answer is found, the box expires, or two independent paths dead-end.
-
At the box's midpoint, check whether a decision is reachable; if not, narrow the question rather than widening the search.
-
When the box expires without an answer, report what was learned and recommend the next bounded step, never an unbounded continuation.
-
If the answer is "this needs to be built", stop: that is a new scoped deliverable, not more research.
Pitfalls
-
"Looking into" something until the deadline with nothing to show.
-
A time box with no defined output, so stopping still produces nothing usable.
-
Extending the box "just a bit more" because the answer feels close; that is the sunk-cost trap.
-
Reporting a pile of observations with no recommendation or decision.
-
Letting the investigation branch into three unrelated questions.
Verification
grep -nE 'stop condition|time box' spike.md # box and stop condition set before start
Report the box, what the investigation found, and the single decision it enables.
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.