复习技巧
- 作者仓库星标 0
- 作者仓库 skills-registry
- 领域
- 工程开发
- 作者声明的智能体
- 未找到明确声明;不据此推断已兼容或已测试
- 静态检查
- 88 / 100 · 启发式扫描,不代表运行安全
- 运行环境证据
- 未执行;当前没有隔离运行测试证据
- 作者 / 版本 / 许可证
- @tomevault-io · 未声明许可证
- 上下文用量评级
- 低消耗
- 接入复杂程度
- 需简单配置
- 是否需要外部接口密钥
- 不需要
- 兼容的系统
- Windows
- 运行环境要求
- 未检出明确要求
- 检测到的文件与系统行为
-
- 只读
- 允许写入 / 修改
- 命令执行
- 网络行为
- 仅限本地
- 安装命令数
- 无(仅作为资料)
需要注意: 未声明工具权限范围;实际权限由运行环境决定。
# Find the real signature
grep -rn "public.*MethodName" src/
# Reviewing and Updating Skills
## When to review
Review the relevant skills and instructions after any work that changes:
- **Public API surface** — new methods, renamed parameters, changed signatures
- **Build or test infrastructure** — new solution files, changed filters, new TFMs
- **Architecture** — new projects, moved files, changed conventions
- **Code generation** — new keywords, changed handler behaviour, new CLI options
- **Documentation pipeline** — new build steps, changed tool behaviour
The rule is simple: if the code changed, check whether the instructions still match.
## What to check
For each skill touched by your changes, verify these in order:
### 1. Code examples compile and are accurate
Every code block in a skill should use real types, real method signatures, and real parameter names from the current codebase. The most common failure mode is **fabricated or stale API usage** — a method that was renamed, a parameter that was added, or a constructor whose signature changed.
To verify, pick each code example and check the actual source:
```powershell
# Find the real signature
grep -rn "public.*MethodName" src/
```
Common drift patterns:
- Constructor gains a new required parameter (skill example silently omits it)
- Method renamed but skill still uses the old name
- Enum values or constants changed
- File paths moved (e.g., `Common/src/` vs `src/Common/`)
### 2. Numeric values and thresholds match source
When a skill states a specific number (priority value, buffer threshold, step count, character limit), verify it against the source constant or definition. These drift silently when someone changes a constant without updating the documentation.
### 3. Scope boundaries are still correct
Check the `USE FOR` / `DO NOT USE FOR` fields in the YAML frontmatter. After adding a new feature or project, an existing skill's scope may need updating — either to include the new area or to explicitly redirect to a new skill.
### 4. Cross-references point to the right places
Each skill's `## Cross-References` section should link to skills that still exist and still cover the referenced topic. If a skill was renamed, split, or merged, update all inbound references.
### 5. No duplication with main instructions
The main `copilot-instructions.md` should contain brief summaries with cross-references to skills for depth. If you find the same detailed content in both places, condense the main instructions copy to a summary and cross-reference.
## Adding a new skill
When a new area of the codebase deserves its own skill:
1. **Create the directory and file:** `.github/skills/<name>/SKILL.md`
2. **Write the YAML frontmatter** with `name`, `description`, `USE FOR`, and `DO NOT USE FOR`
3. **Structure the content:** overview → code examples → configuration tables → common pitfalls → cross-references
4. **Add cross-references** from related skills back to the new one
5. **Update the skill inventory table** in `copilot-instructions.md`
6. **Update the code sample catalog:**
```powershell
.\docs\update-code-sample-catalog.ps1 -UpdateFile .github/skills/<name>/SKILL.md
```
A skill earns its place when the assistant repeatedly struggles with a specific area, not when an area merely exists.
## Root-causing trigger failures
When a skill or instruction existed but you did not follow it, treat that as a defect in the instructions — not just a one-off mistake. The instruction failed to trigger at the right moment, and the fix is to understand why and close the gap.
Ask these questions:
1. **Was the instruction framed for the wrong trigger?** An instruction that says "when you edit a documentation file" may not fire when the task is "fix the CI build." If the instruction applies in both situations, rewrite it to cover both entry points.
2. **Was the instruction buried inside a larger workflow?** A critical step hidden as step 5 of a 5-step process gets skipped when you jump to the end. Promote it to a standalone gate with its own heading.
3. **Did the instruction assume proactive compliance?** Instructions that only describe the happy path ("do X when you change Y") need a reactive counterpart ("if you forgot to do X, here is how to recover — and it still requires doing X").
4. **Did multi-turn conversation obscure the trigger?** When changes accumulate across many conversation turns before a commit, per-edit triggers get lost. Anchor the instruction to the commit point instead, since that is where all changes converge regardless of how they were made.
After root-causing, update the instruction to close the gap — then verify the updated instruction would have caught the original failure.
## Running a full review
For a periodic audit of all skills (e.g., after a major release):
1. **List all skills** and check each against the current source
2. **Use parallel explore agents** to verify 4-5 skills each — look for fabricated APIs, wrong parameters, stale paths, missing examples
3. **Cross-check findings against source code** before fixing — review agents produce false positives (they may flag correct code as wrong)
4. **Track findings** in a structured format (SQL table or similar) with severity and status
5. **Fix in priority order:** critical inaccuracies → missing examples → duplication → cosmetic
### Common false positives
Review agents frequently flag things that are actually correct:
- Generated type APIs that don't appear in hand-written source (they're emitted by the code generator)
- Methods found only via generic type inference (the agent's grep misses them)
- Shared source files in `src-v4/` that are referenced by V5 projects via project references
Always verify against the actual source before changing a skill.
## Design principles for skill content
These principles keep skills effective as AI context:
- **Code-first** — every skill should have at least one copy-paste-ready code example with real syntax
- **Concrete over abstract** — real file paths, real type names, real method signatures
- **Self-contained sections** — each section should make sense without reading the whole skill
- **Tables for reference data** — configuration options, threshold values, priority levels
- **Pitfalls earn their place** — only document pitfalls that have actually caused problems
## Cross-References
- For verifying code samples in documentation, see the "Documentation Code Sample Verification" section in `copilot-instructions.md`
- For the code sample catalog tools, see `docs/CodeSampleCatalog.md`
---
> Source: [corvus-dotnet/Corvus.JsonSchema](https://github.com/corvus-dotnet/Corvus.JsonSchema) — distributed by [TomeVault](https://tomevault.io).
<!-- tomevault:4.0:skill_md:2026-05-23 -->
章节与文件索引
按说明文件中的顺序排列。
何时审查
要检查什么
1. 代码示例编译且准确
2. 数值和阈值与源匹配
3.范围边界仍然正确
文件/命令 -> Common/src/ · src/Common/ · USE FOR · DO NOT USE FOR · ## Cross-References · copilot-instructions.md · .github/skills/<name>/SKILL.md · name
内容校验值 -> 61a08e6f334f
当前原文的自动摘译,尚未人工审校。
何时审查
进行任何更改后,请查看相关技能和说明: 公共 API 界面 — 新方法、重命名参数、更改签名 构建或测试基础设施 — 新解决方案文件、更改过滤器、新 TFM 架构 — 新项目、移动文件、更改约定 代码生成 — 新关键字、更改处理程序行为、新 CLI 选项
要检查什么
对于您的更改涉及的每项技能,请按顺序验证这些技能:
1. 代码示例编译且准确
技能中的每个代码块都应使用当前代码库中的真实类型、真实方法签名和真实参数名称。最常见的故障模式是捏造或陈旧的 API 使用——方法被重命名、参数被添加或构造函数的签名被更改。要进行验证,请选择每个代码示例并检查实际来源:常见漂移模式:构造函数获得新的必需参数(技能示例默默地忽略它)
2. 数值和阈值与源匹配
当技能指定特定数字(优先级值、缓冲区阈值、步数、字符限制)时,根据源常量或定义对其进行验证。当有人在不更新文档的情况下更改常量时,这些会默默地漂移。
3.范围边界仍然正确
检查 YAML frontmatter 中的
USE FOR/DO NOT USE FOR字段。添加新功能或项目后,现有技能的范围可能需要更新 - 要么包含新领域,要么明确重定向到新技能。