メニュー
Use when summarizing Buy vs Build benchmark cases, behavior measurements, expected impact, release notes, or evidence that the ruleset reduces unnecessary code, dependencies, cost, or ownership risk.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Use when evaluating whether to adopt a specific open-source or commercial dependency (npm/PyPI/etc. package or library) — to research its fit, maintenance and health, security and supply-chain risk, license, and footprint before taking it on.
Use when reviewing a code diff, pull request, or local changes for buy-vs-build mistakes, avoidable custom code, unnecessary dependencies, missed built-ins, missed platform features, or weak decision notes.
Use when choosing between options at the same ownership rung — a transport, API style, data store, job/async mechanism, integration pattern, or cache — and you want the one that fits the requirement's real shape (directionality, volume, latency, consistency, failure mode) rather than the most powerful, popular, or familiar default.
Use when coding, designing an implementation, choosing dependencies, adding features, reviewing diffs for custom code, evaluating whether to use built-ins, standard libraries, native platform features, installed packages, open source libraries, commercial services, or an in-house implementation.
Use when a non-obvious buy-vs-build decision was just made and should be recorded durably, or when a project keeps a docs/decisions/ log of Architecture Decision Records (ADRs) for build/buy/reuse and dependency choices.
Use when choosing between built-ins, standard libraries, native platform features, installed packages, open source libraries, commercial products, or in-house implementation for a feature or technical design.
| name | buy-vs-build-gain |
| description | Use when summarizing Buy vs Build benchmark cases, behavior measurements, expected impact, release notes, or evidence that the ruleset reduces unnecessary code, dependencies, cost, or ownership risk. |
Summarize impact from benchmark data without overstating it. Distinguish measured local overhead from behavior benchmarks that still need agent runs.
benchmarks/behavior-cases.json for behavioral test cases.npm run benchmark for local instruction and hook overhead.Use this shape:
Measured
- <metric>: <value>
Behavior cases
- <case>: expected rung <rung>; baseline risk <risk>
Limitations
- <what has not been measured yet>
Never claim the ruleset reduces LOC, cost, or time until a baseline-vs-enabled agent run has actually been measured.