メニュー
Comprehensive verification pattern for Claude Code sessions — build, types, lint, tests, security, and diff review.
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 designing memory or persistent context for an agent — deciding what to persist vs. recompute, how to retrieve only what's relevant, and how to summarize/compact history to fit the context window. Use when an agent forgets, bloats its context, or repeats work.
Interactive installer for Rosetta. Guides users through selecting and installing skills and language-specific rules at user-level or project-level, verifies paths, and optionally optimizes installed files.
Instinct-based learning system that observes sessions via hooks, creates atomic instincts with confidence scoring, and evolves them into skills, commands, and agents.
Use when building or reviewing an evaluation for an LLM feature — assembling a representative test set, choosing pass criteria (exact match, programmatic checks, rubric, or LLM-as-judge), and catching regressions. Use when asking "how do I know this prompt or model change is better?"
Use when writing or reviewing a prompt for an LLM — making it specific, structured, and testable (clear task, role/context, examples, an explicit output contract, decomposition, room to reason) and avoiding the anti-patterns that make prompts brittle. Provider-agnostic.
Pattern for progressively refining context retrieval to solve the subagent context problem.
| name | verification-loop |
| description | Comprehensive verification pattern for Claude Code sessions — build, types, lint, tests, security, and diff review. |
Run after completing a feature, before a PR, after refactoring, or any time you want a full quality check.
Run in order. Stop at build failure — the rest is meaningless if the build is broken.
Build — determine the project's build command and run it. If it fails, stop and fix.
Type check — run the project's type checker (if applicable). Report all errors; fix critical ones before continuing.
Lint — run the linter. Report warnings and errors.
Tests — run the full test suite with coverage. Report pass/fail count and coverage percentage. Target: 80% minimum.
Security — scan for hardcoded secrets, credentials, and debug logging left in source. Use whatever static analysis tools the project has configured; fall back to grep patterns.
Diff review — git diff --stat and git diff HEAD~1 --name-only. Review each changed file for unintended changes, missing error handling, and potential edge cases.
After all phases, report each phase's status (pass/fail with counts), list issues to fix, and give an overall verdict: ready for PR or not.
