菜单
Author tests that match the repo's stack and existing test style, at the cheapest level that catches the regression. Use when the user says "write tests for", "add test coverage", "test this function/handler/component", "this needs tests", or after implementing a feature that lacks tests.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/gnomeria/usbtree --skill write-tests命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
GitHub Actions CI pipeline conventions — job shape, caching, permissions, and action pinning. Use when asked to "set up CI", "add a GitHub Actions workflow", "fix the pipeline", or when creating/reviewing files under .github/workflows/ for Go or Node/TS projects.
Systematic root-cause debugging — reproduce, isolate, fix at the source, prove the fix. Use when the user reports a bug, an error, a crash, flaky behavior, "this doesn't work", "why is this failing", or pastes a stack trace or failing test output.
Set up and maintain a lightweight, file-based knowledge graph of the repo — entities, typed relations, decisions, gotchas — so agents load context fast instead of re-exploring the codebase every session. Use when the user says "knowledge graph", "remember this", "codebase memory", "the agent keeps re-discovering things", or a `.knowledge/` directory exists in the repo.
基于 SOC 职业分类
正在显示 SKILL.md
| name | write-tests |
| description | Author tests that match the repo's stack and existing test style, at the cheapest level that catches the regression. Use when the user says "write tests for", "add test coverage", "test this function/handler/component", "this needs tests", or after implementing a feature that lacks tests. |
Test fails when behavior breaks, passes when it doesn't — nothing more. Run at end, report honestly.
#[cfg(test)] mod tests at bottom, use super::*.dev() in src/usb.rs tests) — reuse existing helper, extend it, don't build parallel one.assert_eq!/assert!. No assertion crates, no mock frameworks.#[rustfmt::skip] + comment per chunk (existing pattern).Cheapest level that catches the regression:
#[test] — pure logic: tree flatten/fold, parsing (usb.ids, sysfs values, descriptors), diff. Default. No hardware, no I/O../target/release/usbtree --demo --dump: exercises scan→tree→render path, no hardware. For wiring bugs, not logic bugs.--demo in pty for interaction (keys, fold, filter). Manual/scripted, last resort — most TUI logic is testable as pure functions on App state; prefer extracting that.Pure parse bug tested through pty = wrong. Keybinding wiring tested as unit = also wrong. Match level to where bug lives.
max_power_parses_sysfs, filter_finds_devices_in_collapsed_hubs — not test1.tempfile::TempDir if already a dep, else skip fs test and test the parse fn on a &str instead — push I/O to edge, test the pure part.cargo test name_fragment first, then full cargo test.