网有NETHAVE
AGENT SKILL

sandbox-test

Test a change to the .asf.yaml parser against a real GitHub repository, using the "Test a change against the sandbox repo" workflow. Covers one pull request, a local branch, or a batch of open PRs. Builds a set/update/unset test matrix, runs it noop then live, verifies every claim by reading the GitHub API back, resets the sandbox, and then proposes a review comment and a disposition for the human to approve. Invoke when asked to "sandbox-test this", "test PR NNN against the sandbox", "does this actually work on GitHub", "test the open PRs", or before approving or merging any PR that touches asfyaml/feature/.

使用场景测试与质量
License 许可
Apache-2.0
最近更新
2026-11-09
Stars
★ 16
Forks
29

01 前言 Front Matter

name
review-checklist
description
Generates a language-specific code review checklist tailored to the files being changed.
argument-hint
[language-or-file]

03 SKILL.md 正文

Review Checklist Skill

This skill generates targeted code review checklists customized for the specific programming language or files in your change.

When to Use

  • Before submitting a PR for review
  • When reviewing code from teammates
  • To ensure consistency with language best practices
  • To catch common language-specific pitfalls

Supported Languages

  • TypeScript — Type safety, async/await patterns, React/Node conventions
  • C# — Null safety, async patterns, LINQ, .NET framework usage
  • C++ — Memory management, smart pointers, const correctness, thread safety
  • Python — Type hints, exception handling, PEP 8, async patterns
  • Go — Error handling, goroutines, channels, interfaces
  • Java — NullPointerException handling, resource management, collection APIs

Checklist Categories

Each checklist includes checks for:

Type Safety

  • Type annotations present and correct
  • No unsafe casts or type assertions
  • Proper generic type usage
  • Runtime type validation where needed

Error Handling

  • All exceptions/errors properly handled
  • No silent failures or swallowed exceptions
  • Appropriate error recovery strategies
  • User-facing error messages are clear

Testing & Coverage

  • Unit tests for new logic
  • Edge cases covered
  • Error paths tested
  • Integration points validated

Security

  • Input validation and sanitization
  • No hardcoded credentials or secrets
  • Proper authentication/authorization checks
  • SQL injection and XSS prevention

Documentation

  • Public API methods documented
  • Complex logic has explanatory comments
  • Breaking changes noted
  • Examples provided for non-obvious behavior

Output Format

Checklists are formatted as markdown with categories as sections, each containing checkboxes:

## [Language] Code Review Checklist

### Type Safety
- [ ] All variables have explicit types
- [ ] No unnecessary type assertions
- ...

### Error Handling
- [ ] Exception handling is explicit
- [ ] Error messages are helpful
- ...

### Testing & Coverage
- [ ] New code has unit tests
- [ ] Edge cases are covered
- ...

### Security
- [ ] Input validation present
- [ ] No credential leaks
- ...

### Documentation
- [ ] Public APIs are documented
- [ ] Complex logic is explained
- ...

Example Usage

/review-checklist TypeScript
/review-checklist C#
/review-checklist for src/auth.ts
/review-checklist Python

The skill automatically detects the language from file extensions when provided, or uses the language name directly to generate the appropriate checklist.