skill-management
Discover, activate, and manage agent skills following the Skill Protocol
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
메뉴
Discover, activate, and manage agent skills following the Skill Protocol
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Health check procedures D1–D14 for the Audit agent — structural validation, attention budget, version checks, workspace integrity, and static audit
Configure and manage Model Context Protocol servers for external tool access
Review a UI for accessibility — WCAG 2.1 AA compliance, semantic HTML, ARIA usage, keyboard navigation, focus management, colour contrast, and screen reader compatibility
Design or review a REST or GraphQL API — resource modeling, versioning strategy, error contract, OpenAPI/schema-first workflow, and security baseline
Generate a CHANGELOG.md entry from staged changes, a commit range, or a PR diff — following Keep a Changelog format with conventional commit classification
Set up and audit environment variable management — create .env.example, add startup validation, separate secrets from config, and document every variable
| name | skill-management |
| description | Discover, activate, and manage agent skills following the Skill Protocol |
| compatibility | >=1.4 |
Skill metadata: version "1.1"; license MIT; tags [skills, workflow, discovery, management]; compatibility ">=1.4"; recommended tools [codebase, editFiles, fetch].
Skills are reusable markdown-based behavioural instructions that teach the agent how to perform a workflow. Unlike tools (§11) which are executable scripts, skills are declarative. They follow the Agent Skills open standard: SKILL.md with YAML frontmatter (name, description) plus step-by-step instructions.
Use description as the discovery surface. Do not rely on custom keys like stacks; VS Code ignores them.
Skills are loaded on demand — the agent reads a skill's SKILL.md only when the description field matches the current task context. Do not pre-load all skills.
Task requires a workflow
│
├─ 1. SCAN — check local skill directories
│ Locations: .github/skills/, .claude/skills/, .agents/skills/
│ Also: installed agent plugins and extension-contributed skills
│ ├─ Match found → READ the full SKILL.md, follow its instructions
│ └─ No match → ↓
│
├─ 2. SEARCH (if enabled by skill search preference setting)
│ ├─ Search official repos (anthropics/skills, github/awesome-copilot) THEN:
│ │ community sources (GitHub search, awesome-agent-skills)
│ │ ├─ Found → evaluate fit, quality-check, adapt, save locally
│ │ └─ Not found → ↓
│
└─ 3. CREATE — author a new skill from scratch
- Save to .github/skills/<kebab-name>/SKILL.md
| Priority | Location | Scope |
|---|---|---|
| 1 (highest) | .github/skills/<name>/SKILL.md | Project — checked into version control |
| 2 | .claude/skills/<name>/SKILL.md, .agents/skills/<name>/SKILL.md | Project (alt) — Claude/Agent-format directories |
| 3 | ~/.copilot/skills/<name>/SKILL.md | Personal — shared across all projects for one user |
| 4 | ~/.claude/skills/<name>/SKILL.md, ~/.agents/skills/<name>/SKILL.md | Personal (alt) — alternative personal paths |
| 5 | Agent plugins (@agentPlugins) | Plugin — installed via Extensions view (VS Code 1.110+) |
| 6 | Extension chatSkills contribution | Extension — VS Code extensions contributing skills via package.json |
| 7 | Organization-level agents | Org — published at GitHub org level for all members |
Custom paths: Use
chat.skillsLocationsto add custom directories. Keep legacychat.agentSkillsLocationsin sync for older VS Code builds.
| Setting | / menu | Auto-load | Use case |
|---|---|---|---|
| Default | Yes | Yes | General-purpose |
user-invocable: false | No | Yes | Background knowledge |
disable-model-invocation: true | Yes | No | On-demand only |
| Both set | No | No | Disabled |
Subagents inherit this protocol. A subagent may read and follow any skill but must flag new skill creation to the parent agent for confirmation.