一键导入
verification-loop
Comprehensive verification system for Claude Code sessions. Use after implementation to validate correctness.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Comprehensive verification system for Claude Code sessions. Use after implementation to validate correctness.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Web取得の標準前処理レイヤー。URL・公式ドキュメント・ブログ・ニュース・OSSページを読むときは、原則として毎回このSkillでDefuddleを使い本文をMarkdown/JSON化してから読む。生HTMLのまま要約・分析・比較・レビューしない。トリガー例: URLを読む/Webページ要約/OSS調査/公式Doc確認/記事解析/競合サイト確認/Web一次情報確認。
Formal evaluation framework for Claude Code sessions implementing eval-driven development (EDD) principles.
Frontend development patterns for React, Next.js, state management, performance optimization, and UI best practices.
Example project-specific skill template. Use as a starting point when creating guidelines for your own projects.
Practical prompt patterns and techniques for effective Claude Code usage. Reference when crafting prompts, planning workflows, or debugging interactions.
Use this skill when the user explicitly asks for TDD or a tests-first workflow, or when developing a new feature with a test coverage requirement. Provides a structured red-green-refactor workflow with unit, integration, and E2E test guidance.
| name | verification-loop |
| description | Comprehensive verification system for Claude Code sessions. Use after implementation to validate correctness. |
| when_to_use | Use when implementation is complete and the user wants verification before handoff, commit, or PR creation. |
A comprehensive verification system for Claude Code sessions.
Invoke this skill:
Pick the right command for each phase from the project's own configuration (package.json scripts, Makefile, CI workflow, etc.). The commands below are examples, not prescriptions. Summarize long output instead of pasting it verbatim.
Check that the project builds. Examples: npm run build, cargo check, go build ./..., shellcheck -S warning *.sh.
If the build fails, STOP and fix before continuing.
Run the project's type checker if it has one. Examples: npx tsc --noEmit (TypeScript), pyright . or mypy . (Python).
Report all type errors. Fix critical ones before continuing.
Run the project's linter. Examples: npm run lint, ruff check ., shellcheck.
Run the project's test suite, with coverage if available. Examples: npm test -- --coverage, pytest --cov, bash tests/run-unit-tests.sh.
Report:
Check for leaked secrets in the changes:
git diff for credentials, tokens, and keys. Do not report PASS based on a repository-wide grep.Also check for leftover debug output (e.g., console.log, print) in the changed files.
# Show what changed
git diff --stat
git diff HEAD~1 --name-only
Review each changed file for:
After running all phases, produce a verification report:
VERIFICATION REPORT
==================
Build: [PASS/FAIL]
Types: [PASS/FAIL] (X errors)
Lint: [PASS/FAIL] (X warnings)
Tests: [PASS/FAIL] (X/Y passed, Z% coverage)
Security: [PASS/FAIL] (X issues)
Diff: [X files changed]
Overall: [READY/NOT READY] for PR
Issues to Fix:
1. ...
2. ...
This skill complements PostToolUse hooks but provides deeper verification. Hooks catch issues immediately; this skill provides comprehensive review. If you need periodic automated re-verification, use a hook or the /loop command; instructions in this skill body cannot schedule themselves.