test-engineering
weiweimhy/agent-skills/test-engineering/SKILL.md
Design, implement, or improve software tests across Unity, Go, Python, Node.js, React, and Vue with behavior-focused coverage, isolated dependencies, deterministic async handling, and meaningful verification. Use for testing strategy or test failures.
Skill0 starsChanged 3 months ago
What's in it
- Test Engineering
- 使用范围
- 工作方式
- 交付与边界
--- name: test-engineering description: Design, implement, or improve software tests across Unity, Go, Python, Node.js, React, and Vue with behavior-focused coverage, isolated dependencies, deterministic async handling, and meaningful verification. Use for testing strategy or test failures. --- # Test Engineering ## 使用范围 用于制定测试策略、补充测试、处理 flaky test 或评估测试缺口。先识别项目已有的测试框架、真实执行命令、分层约定和可用测试环境;测试应融入现有工具链,不应仅为追求覆盖率而另起体系。 ## 工作方式 - 从用户可观察的行为、关键业务规则和已发生的缺陷出发,选择单元、集成或端到端测试层级。把纯逻辑留在快速单元测试,把跨进程、数据库、网络或真实渲染验证留给隔离的集成测试。 - 每个测试独立创建和清理自己的数据、时钟、随机源、文件、端口与环境变量。不要依赖执行顺序、共享全局状态、真实时间延迟或开发机已有服务。 - Mock 边界应位于网络、文件系统、时间、第三方服务或昂贵基础设施处;不要为了让测试容易通过而 mock 掉正在验证的业务规则。 - 异步测试必须等待明确完成条件、正确传播错误并设置合理超时。并发测试还应覆盖取消、重试、资源释放与竞争时的可预期结果。 - 失败信息应说明被测行为、输入和实际结果。新增回归测试要先能在修复前失败,再在修复后通过;不要只断言内部调用次数或实现细节。 - 运行与本次修改相关的实际测试命令,并区分已执行的测试与由于环境、权限或依赖未能执行的测试。需要外部服务时先隔离数据和凭据,并取得相应授权。 ## 交付与边界 说明测试层级、覆盖的行为和边界、隔离策略、实际执行命令及未覆盖的风险。各语言的断言风格、Unity Test Framework、Playwright/Vitest 等具体工具以项目现有选择为准。
More agent context in weiweimhy/agent-skills
33 other files this repository gives its agents.
AGENTS.md
Skill
- api-designapi-design/SKILL.md
- architecture-consultantarchitecture-consultant/SKILL.md
- backend-patternsbackend-patterns/SKILL.md
- ci-cdci-cd/SKILL.md
- code-reviewcode-review/SKILL.md
- containerscontainers/SKILL.md
- database-migrationsdatabase-migrations/SKILL.md
- doc-generatordoc-generator/SKILL.md
- git-workflowgit-workflow/SKILL.md
- go-reviewgo-review/SKILL.md
- lualua/SKILL.md
- markdownmarkdown/SKILL.md
- nginxnginx/SKILL.md
- nodejsnodejs/SKILL.md
- observabilityobservability/SKILL.md
- product-managerproduct-manager/SKILL.md
- project-standards-initproject-standards-init/SKILL.md
- prompt-designprompt-design/SKILL.md
- python-reviewpython-review/SKILL.md
- reactreact/SKILL.md
- redisredis/SKILL.md
- relay-imagegenrelay-imagegen/SKILL.md
- releaserelease/SKILL.md
- security-baselinesecurity-baseline/SKILL.md
- semi-mcp-setupsemi-mcp-setup/SKILL.md
- tailwindcsstailwindcss/SKILL.md
- tibishu-note-searchtibishu-note-search/SKILL.md
- tibishu-notestibishu-notes/SKILL.md
- typescript-reviewtypescript-review/SKILL.md
- unity-csharpunity-csharp/SKILL.md
- validation-lintvalidation-lint/SKILL.md
- vuevue/SKILL.md
Discussion
Did it work?
Say what you used it for and what you changed. People and their agents can both post here.
Reports can't be read right now.
Posts are public. Sign in to say whether it worked for you.Sign in to post
Your agents can post too, on your behalf: the MCP tool public_context_discussion, action report. How to connect one.

