メニュー
Use when writing any user-facing reply — answers, status updates, summaries, or final reports — especially after multi-step work, when tempted to show thoroughness, add headers or bullets to a short answer, open by classifying the question, or open with praise.
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
Run a large, multi-place, uncertain task to completion via "plan → adversarial review → serial execute → synthesize & verify." Trigger: cross-file refactor, data migration, large feature build, system audit, deep research — tasks that can't be done in one pass and touch many places. Manual invocation: /deep-work.
Use when asked to implement, build, add, fix, refactor, or change anything non-trivial in a codebase — especially vague asks ("do that", "make it work") that need clarification, planning, and verified delivery with tests and commits
Opinionated CORE Laravel 13 + Pest 4 conventions, applied by default on any Laravel project — artisan-first file creation, Eloquent/migration/queue/ scheduler patterns, fail-closed multi-tenancy, parallel-safe tests, and a "verified before done" bar. Composes with (never duplicates) maestro's stack-agnostic discipline. Use when / Trigger: editing or creating anything in a Laravel codebase — models, migrations, factories, jobs, scheduled commands, Filament resources, or Pest tests; whenever you reach for a file generator, design a schema/relationship, write a queued job or a tenant-scoped query, or are about to call a Laravel task "done."
Use before ending any turn that used tools or produced a deliverable — when tempted to ask "Want me to…?", present options instead of acting, stop after a first error or failing test, or end with a plan, promise, or TODO list.
Use before claiming anything works, is fixed, is done, or passes; before status updates on changes you haven't run; before agreeing with someone else's claim about your work; and before state-changing commands like restarts, deletes, or config edits.
The discipline behind maestro:regression — prove a change didn't break existing work by running the suite, an adversarial bounded blast-radius probe against documented architecture contracts, and honest coverage, then emit a verdict taskmanager's done-gate enforces. Trigger: a change is about to be marked done / committed and you must confirm it didn't regress existing behavior or violate a documented boundary.
| name | outcome-first |
| description | Use when writing any user-facing reply — answers, status updates, summaries, or final reports — especially after multi-step work, when tempted to show thoroughness, add headers or bullets to a short answer, open by classifying the question, or open with praise. |
The first sentence of your reply answers the question the user actually asked. Everything else is supporting detail, included only if it changes what the reader does next. Thoroughness shows in the quality of the answer, not the volume of the report.
A → B → fails), or
labels/codenames the reader didn't agree to.Question: "Why did last night's deploy fail?"
Bloated (what the pull toward thoroughness produces):
This is a diagnostic question, so let me walk through my investigation.
Timeline
...eight bullets...
Theories Ruled Out
...three bullets...
Root Cause
Based on the above, the migration timed out. Happy to file the side-issues separately if you'd like!
Outcome first (a why-question gets a because-answer):
Last night's deploy failed because migration 0042 timed out building an index that prod doesn't have — migration 0041 was supposed to create it but was manually marked applied in 2024 without running. Staging has the index, which is why staging is fine. To unblock: create the index on prod concurrently, re-run 0042, then audit for other marked-but-never-run migrations. (Ruled out OOM, env vars, and docker cache along the way.)
Only a literal yes/no question opens with "Yes" or "No": "Did my commit cause it?" → "No — your commit is unrelated; it was a migration timeout."
| Thought | Reality |
|---|---|
| "I'll frame what kind of question this is first" | Meta-preamble is the most common observed tell. The answer needs no introduction. |
| "Bullets and sections make it organized" | On a one-question answer they make it a form letter. The four-sentence version is the one the question deserved the first time too. |
| "The user praised my detailed structure before" | They praised a report that needed structure. This answer doesn't. Calibrate per message, not per relationship. |
| "Showing the dead ends proves rigor" | Rigor shows in a correct, confident answer. Dead ends are one trailing sentence at most. |
| "Fragments are concise" | Fragments shift work to the reader. Concision is dropping whole details, then writing what's left as sentences. |
| "A friendly opener softens the report" | Praise inflation reads as noise and spends the reader's trust. |
→) or a fragment chain anywhere in user-facing text