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/KingInYellows/yellow-plugins --skill test-skill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。
正在显示 SKILL.md
| name | test-skill |
| description | Skill fixture used to validate the skill-reference resolver in lint-plugins.sh. |
| user-invocable | false |
Body content. The body intentionally contains a stray name: line below to
regression-test the P1 fix from PR #261, which scoped name extraction to
frontmatter so body prose can no longer corrupt the known-skills set.
The next line deliberately starts at column 1 with no blockquote or code
fence prefix, so an awk extractor running on the whole file (the pre-fix
bug) would match ^name: against it. The lint script must NOT treat this
body-prose name: as a skill — only the frontmatter name: test-skill
should index.
name: should-not-be-collected
Expert guidance for creating Claude Code skills and agents. Use when working with SKILL.md files, authoring new skills, creating slash commands, or designing agent workflows.
Run 1-3 parallel quick-audit passes (code quality, security, error handling) over a diff and gate on findings. Use when smart-submit or gt-amend need to audit uncommitted changes before committing.
Decompose a feature into stacked PRs, ordered by dependency (plan-only). Use when breaking a feature into reviewable stacked PRs.