agentsclimarketplace

Ccdawn bug review

Skill CCDawn/codex-skills/skills/engineering/ccdawn-bug-review

Chinese-first Codex skill pack for intent alignment, dynamic routing, lightweight development, code review, UI design, and AI research workflows.

Install
npx -y skills add CCDawn/codex-skills --skill ccdawn-bug-review

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 4 stars4 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Use when an observed bug, correctness regression, failing test, build failure, abnormal behavior, or unresolved root cause needs Chinese-first diagnosis and bounded repair; obvious local inefficiency found during feature work stays with its owner, while measurable performance work uses ccdawn-performance-engineering.

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

4.1 KB, as published. Nobody here has run it

CCDawn Bug Review

目标

直接承接 bug 诊断与契约内修复：先用证据定位根因，再做最小修复和风险相称的验证。不要求展示固定阶段、长检查表或完整思维过程。

BRT interface

  • Context Boundary: 预期行为、实际症状、失败命令/日志、相关代码与测试、允许修改面和非目标。
  • Output Contract: 根因状态、证据链、最小修复或下一 probe、影响范围、验证结果和剩余风险。
  • Allowed Action: 用户要求修复且边界清楚时可直接读取、复现、修改和验证；只要求审查时保持只读。
  • Success Evidence: 失败可复现或有等价证据，根因与修复存在因果联系，目标验证通过且未越界。
  • Stop Condition: 缺少必要对象/权限、根因仍不稳定且写入会扩大误改、需要破坏性动作、需求冲突或风险越过当前授权。
  • Route Out: 契约内修复、root-cause-tracing、ccdawn-performance-engineering、ccdawn-planning、ccdawn-development-cleanup、ccdawn-pr-review、ccdawn-brt 或 BLOCKED。

统一调用契约

  • 只处理 BRT interface 范围；不匹配时回 ccdawn-brt 或更具体 owner，复合任务不吞其他 owner。
  • 用户可见内容默认中文，完成只报状态、产出、证据和剩余风险；代码、命令、路径、错误原文、API/协议、skill 名和枚举保留原样；Route Out 仅以 BRT interface 为准，末行写 下一步建议: <一个具体动作>。

调试契约

  1. 写清 Expected / Actual / Scope，先读本地可得的代码、日志、失败测试和近期 diff。
  2. 优先复用现有失败测试或稳定复现；无法直接运行时，用调用链、状态变化或日志建立最窄证据链。
  3. 定位 owning surface，沿数据流或控制流追到最早错误来源。来源藏在深层链路时才加载 root-cause-tracing。
  4. 可直接确认的 N+1、循环 I/O、重复计算等明显低效留给当前 owner；已观察故障、正确性回归或根因不明才由本 skill 诊断；可测性能问题路由 ccdawn-performance-engineering。
  5. 标记根因：CONFIRMED / HYPOTHESIS / ENVIRONMENT / TEST_ISSUE / REQUIREMENT_MISMATCH。只有 CONFIRMED 才进入行为修复；其他状态继续低风险 probe 或路由正确 owner。
  6. 用户已授权修复且根因、边界、回滚和验证清楚时，直接做最小修复；不为流程形式停下确认。
  7. 运行能证明因果关系的窄验证，再按影响面决定是否扩展。区分实现失败、测试意图过期、环境失败和需求不一致。

Bug 测试锚点

Bug 修复不转交 TDD owner。已有失败测试或稳定复现直接作为 RED；若回归风险显著且没有自动化证据，在当前 owner 内补一个最小行为测试，确认它因目标缺陷失败后修复并得到 GREEN。局部、可逆且现有验证足够时直接修复，不为形式制造 RED。

测试只证明已确认的预期行为，不替代根因诊断。测试意图过期时标记 TEST_ISSUE，需求不一致时标记 REQUIREMENT_MISMATCH，不得为了 GREEN 固化旧约束。

流程重量

简单、局部、可逆问题直接修复并验证，不 planning、拆分或创建 worktree。只有真实设计分叉、跨系统迁移或边界不稳才进入 ccdawn-planning；多文件或复杂调用链本身不升级流程。

输出

普通修复只需汇报：根因 / 修改 / 验证 / 剩余风险 / 下一步建议。只读审查或阻塞时再补：

Bug 判断:
- 根因状态:
- 关键证据:
- 影响范围:
- 建议动作:
- 停止条件:
下一步建议: <一个具体动作>

不得为了让测试通过而削弱正确行为，也不得在根因仍是猜测时堆叠补丁。

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.