thun-res/vlink / .agents/skills/deep-review
deep-review
Verified对 VLink 全仓或指定范围执行只读重度代码评审,按功能规划覆盖范围和 subagent,逐文件检查正确性、并发、性能与规范,经对抗复核后给出中文 证据结论。用户要求"全面 review"、"深度检查全仓"、"大规模审计代码" 或"对抗评审"时使用;明确要求落盘时生成报告。
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 1.5 KB |
| agents/openai.yaml | config | 231 B |
| references/AGENT-ORCHESTRATION.md | markdown | 1.5 KB |
| references/REPORTING.md | markdown | 1014 B |
| references/REVIEW-DIMENSIONS.md | markdown | 2.2 KB |
| references/SCOPE-PLANNING.md | markdown | 1.8 KB |
references/REVIEW-DIMENSIONS.md
View on GitHub# 逐文件评审维度
## 通用检查
- **正确性**:边界条件、溢出与截断、未初始化状态、生命周期、错误传播、
资源清理和平台分支。
- **并发**:锁顺序、原子内存序、竞态、TOCTOU、线程退出、MessageLoop
与 Timer 的现有语义。
- **性能**:热路径分配、隐藏拷贝、临时字符串、锁竞争、批处理退化和
缺失的合理分支预测。
- **接口**:参数与返回值、ABI/线格式、配置默认值、后端行为一致性和
Doxygen 与实现的一致性。
- **安全**:命令与路径注入、越界访问、不可信输入边界、敏感信息泄露、
workflow 权限和 secret 使用。
- **工程与发布**:CMake 安装/导出、包元数据、脚本错误传播、workflow
触发条件、产物路径和发布前置条件。
- **克制**:过度封装、防御、缓存、重试、兼容层、死代码或未被当前需求
使用的设计。
重点核对 `Bytes` 拷贝构造/赋值的深拷贝代价、`deep_copy`、
非平凡载荷按值传递以及 `VLIKELY`/`VUNLIKELY` 的合理使用。
## 按文件类型加载规则
- C++:`.agents/languages/CPP.md`;
- 纯 C/C API:`C.md` 与 `CPP.md` 上位约束;
- Shell、Batch、PowerShell、CMake、Python:对应 languages 分册;
- 测试:`.agents/testing/UNIT-TESTS.md`;
- 文档、Wiki、UI: `.agents/DOCS-AND-FORMATTING.md`;
- workflow、Dockerfile 和工程配置: `.agents/CI-AND-PR.md`、
`.agents/REPO-REFERENCE.md` 与相邻文件。
规则没有覆盖的文件类型沿用相邻文件,不得自行引入外部惯例。
## 严重级
- **P0 缺陷**:可触发崩溃、数据破坏、安全问题、死锁或 ABI/线格式破坏,
阻塞合并。
- **P1 风险**:在明确场景产生错误行为、资源泄漏、显著性能回归、
跨平台失败或关键测试缺口。
- **P2 规范**:当前行为未必错误,但违反仓库硬规则,或存在文档、注释、
测试与实现不一致。
严重级始终按实际影响判定;公开契约冲突或关键测试缺口不得因出现在文档、
注释或测试中而自动降为 P2。
每条发现必须包含 `文件:行号`、严重级、缺陷描述、触发条件、影响和最小
修复方向。纯偏好或无触发场景的猜测不列为缺陷。
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |