[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M718AVKT0RG9D0ZXA93X8GJF":3,"skill-same:01M718AVKT0RG9D0ZXA93X8GJF":101},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":23,"publishedAt":44,"linkOnly":45,"markdownContent":46,"markdownBody":47,"frontMatter":48,"refName":50,"commitSha":51,"skillDirectory":52,"forksCount":53,"repositoryHtmlUrl":54,"repositoryHomepageUrl":55,"ownerHtmlUrl":56,"ownerBlogUrl":57,"ownerDescription":6,"updatedAt":58,"install":59,"resources":72},"01M718AVKT0RG9D0ZXA93X8GJF","agent-development",null,"This skill should be used when the user asks to \"create an agent\", \"add an agent\", \"write a subagent\", \"agent frontmatter\", \"when to use description\", \"agent examples\", \"agent tools\", \"agent colors\", \"autonomous agent\", or needs guidance on agent structure, system prompts, triggering conditions, or agent development best practices for Claude Code plugins.","anthropics\u002Fclaude-plugins-official","anthropics","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F76263028?v=4","01M697XGDPXJAVQ8391V2BCYTE","Anthropic","01M7188MR49TZ9N2ZKPXJ7M7HY",46906,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fclaude-plugins-official\u002Fblob\u002Fmain\u002Fplugins\u002Fplugin-dev\u002Fskills\u002Fagent-development\u002FSKILL.md",[18],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43],"development","agent development","guidance","system prompts","examples","best practices","use","system","for","add","Tools","claude code plugins","frontmatter","write","structure","create","subagent","code","when","autonomous","2026-11-10 10:12:30",false,"---\nname: agent-development\ndescription: This skill should be used when the user asks to \"create an agent\", \"add an agent\", \"write a subagent\", \"agent frontmatter\", \"when to use description\", \"agent examples\", \"agent tools\", \"agent colors\", \"autonomous agent\", or needs guidance on agent structure, system prompts, triggering conditions, or agent development best practices for Claude Code plugins.\nversion: 0.1.0\n---\n\n# Agent Development for Claude Code Plugins\n\n## Overview\n\nAgents are autonomous subprocesses that handle complex, multi-step tasks independently. Understanding agent structure, triggering conditions, and system prompt design enables creating powerful autonomous capabilities.\n\n**Key concepts:**\n- Agents are FOR autonomous work, commands are FOR user-initiated actions\n- Markdown file format with YAML frontmatter\n- Triggering via description field with examples\n- System prompt defines agent behavior\n- Model and color customization\n\n## Agent File Structure\n\n### Complete Format\n\n```markdown\n---\nname: agent-identifier\ndescription: Use this agent when [triggering conditions]. Typical triggers include [scenario 1 in prose], [scenario 2 in prose], and [scenario 3 in prose]. See \"When to invoke\" in the agent body for worked scenarios.\nmodel: inherit\ncolor: blue\ntools: [\"Read\", \"Write\", \"Grep\"]\n---\n\nYou are [agent role description]...\n\n## When to invoke\n\n[Two to four representative scenarios written as prose, e.g.:]\n- **[Scenario name].** [What the situation looks like and what the agent should do.]\n- **[Scenario name].** [Same.]\n\n**Your Core Responsibilities:**\n1. [Responsibility 1]\n2. [Responsibility 2]\n\n**Analysis Process:**\n[Step-by-step workflow]\n\n**Output Format:**\n[What to return]\n```\n\n## Frontmatter Fields\n\n### name (required)\n\nAgent identifier used for namespacing and invocation.\n\n**Format:** lowercase, numbers, hyphens only\n**Length:** 3-50 characters\n**Pattern:** Must start and end with alphanumeric\n\n**Good examples:**\n- `code-reviewer`\n- `test-generator`\n- `api-docs-writer`\n- `security-analyzer`\n\n**Bad examples:**\n- `helper` (too generic)\n- `-agent-` (starts\u002Fends with hyphen)\n- `my_agent` (underscores not allowed)\n- `ag` (too short, \u003C 3 chars)\n\n### description (required)\n\nDefines when Claude should trigger this agent. **This is the most critical field** — it is loaded into context whenever the agent is registered, so the harness can decide when to dispatch.\n\n**Must include:**\n1. Triggering conditions (\"Use this agent when...\")\n2. A short prose summary of the typical trigger scenarios\n3. A pointer to a \"When to invoke\" section in the agent body for the detailed worked scenarios\n\n**Format:**\n```\nUse this agent when [conditions]. Typical triggers include [scenario 1 in prose], [scenario 2 in prose], and [scenario 3 in prose]. See \"When to invoke\" in the agent body for worked scenarios.\n```\n\n**Best practices:**\n- Name 2-4 trigger scenarios in the prose summary\n- Cover both proactive (assistant invokes itself) and reactive (user requests) triggering\n- Cover different phrasings of the same intent\n- Be specific about when NOT to use the agent\n- Put detailed scenarios in the body under \"When to invoke\" as a bullet list of prose descriptions\n\n### model (required)\n\nWhich model the agent should use.\n\n**Options:**\n- `inherit` - Use same model as parent (recommended)\n- `sonnet` - Claude Sonnet (balanced)\n- `opus` - Claude Opus (most capable, expensive)\n- `haiku` - Claude Haiku (fast, cheap)\n\n**Recommendation:** Use `inherit` unless agent needs specific model capabilities.\n\n### color (required)\n\nVisual identifier for agent in UI.\n\n**Options:** `blue`, `cyan`, `green`, `yellow`, `magenta`, `red`\n\n**Guidelines:**\n- Choose distinct colors for different agents in same plugin\n- Use consistent colors for similar agent types\n- Blue\u002Fcyan: Analysis, review\n- Green: Success-oriented tasks\n- Yellow: Caution, validation\n- Red: Critical, security\n- Magenta: Creative, generation\n\n### tools (optional)\n\nRestrict agent to specific tools.\n\n**Format:** Array of tool names\n\n```yaml\ntools: [\"Read\", \"Write\", \"Grep\", \"Bash\"]\n```\n\n**Default:** If omitted, agent has access to all tools\n\n**Best practice:** Limit tools to minimum needed (principle of least privilege)\n\n**Common tool sets:**\n- Read-only analysis: `[\"Read\", \"Grep\", \"Glob\"]`\n- Code generation: `[\"Read\", \"Write\", \"Grep\"]`\n- Testing: `[\"Read\", \"Bash\", \"Grep\"]`\n- Full access: Omit field or use `[\"*\"]`\n\n## System Prompt Design\n\nThe markdown body becomes the agent's system prompt. Write in second person, addressing the agent directly.\n\n### Structure\n\n**Standard template:**\n```markdown\nYou are [role] specializing in [domain].\n\n**Your Core Responsibilities:**\n1. [Primary responsibility]\n2. [Secondary responsibility]\n3. [Additional responsibilities...]\n\n**Analysis Process:**\n1. [Step one]\n2. [Step two]\n3. [Step three]\n[...]\n\n**Quality Standards:**\n- [Standard 1]\n- [Standard 2]\n\n**Output Format:**\nProvide results in this format:\n- [What to include]\n- [How to structure]\n\n**Edge Cases:**\nHandle these situations:\n- [Edge case 1]: [How to handle]\n- [Edge case 2]: [How to handle]\n```\n\n### Best Practices\n\n✅ **DO:**\n- Write in second person (\"You are...\", \"You will...\")\n- Be specific about responsibilities\n- Provide step-by-step process\n- Define output format\n- Include quality standards\n- Address edge cases\n- Keep under 10,000 characters\n\n❌ **DON'T:**\n- Write in first person (\"I am...\", \"I will...\")\n- Be vague or generic\n- Omit process steps\n- Leave output format undefined\n- Skip quality guidance\n- Ignore error cases\n\n## Creating Agents\n\n### Method 1: AI-Assisted Generation\n\nUse this prompt pattern (extracted from Claude Code):\n\n```\nCreate an agent configuration based on this request: \"[YOUR DESCRIPTION]\"\n\nRequirements:\n1. Extract core intent and responsibilities\n2. Design expert persona for the domain\n3. Create comprehensive system prompt with:\n   - Clear behavioral boundaries\n   - Specific methodologies\n   - Edge case handling\n   - Output format\n   - A \"When to invoke\" section listing 2-4 trigger scenarios as prose bullets\n4. Create identifier (lowercase, hyphens, 3-50 chars)\n5. Write description with triggering conditions and a short prose summary of trigger scenarios\n\nReturn JSON with:\n{\n  \"identifier\": \"agent-name\",\n  \"whenToUse\": \"Use this agent when... Typical triggers include [...]. See \\\"When to invoke\\\" in the agent body.\",\n  \"systemPrompt\": \"You are...\"\n}\n```\n\nThen convert to agent file format with frontmatter.\n\nSee `examples\u002Fagent-creation-prompt.md` for complete template.\n\n### Method 2: Manual Creation\n\n1. Choose agent identifier (3-50 chars, lowercase, hyphens)\n2. Write description with examples\n3. Select model (usually `inherit`)\n4. Choose color for visual identification\n5. Define tools (if restricting access)\n6. Write system prompt with structure above\n7. Save as `agents\u002Fagent-name.md`\n\n## Validation Rules\n\n### Identifier Validation\n\n```\n✅ Valid: code-reviewer, test-gen, api-analyzer-v2\n❌ Invalid: ag (too short), -start (starts with hyphen), my_agent (underscore)\n```\n\n**Rules:**\n- 3-50 characters\n- Lowercase letters, numbers, hyphens only\n- Must start and end with alphanumeric\n- No underscores, spaces, or special characters\n\n### Description Validation\n\n**Length:** 10-5,000 characters\n**Must include:** Triggering conditions and examples\n**Best:** 200-1,000 characters with 2-4 examples\n\n### System Prompt Validation\n\n**Length:** 20-10,000 characters\n**Best:** 500-3,000 characters\n**Structure:** Clear responsibilities, process, output format\n\n## Agent Organization\n\n### Plugin Agents Directory\n\n```\nplugin-name\u002F\n└── agents\u002F\n    ├── analyzer.md\n    ├── reviewer.md\n    └── generator.md\n```\n\nAll `.md` files in `agents\u002F` are auto-discovered.\n\n### Namespacing\n\nAgents are namespaced automatically:\n- Single plugin: `agent-name`\n- With subdirectories: `plugin:subdir:agent-name`\n\n## Testing Agents\n\n### Test Triggering\n\nCreate test scenarios to verify agent triggers correctly:\n\n1. Write agent with specific triggering examples\n2. Use similar phrasing to examples in test\n3. Check Claude loads the agent\n4. Verify agent provides expected functionality\n\n### Test System Prompt\n\nEnsure system prompt is complete:\n\n1. Give agent typical task\n2. Check it follows process steps\n3. Verify output format is correct\n4. Test edge cases mentioned in prompt\n5. Confirm quality standards are met\n\n## Quick Reference\n\n### Minimal Agent\n\n```markdown\n---\nname: simple-agent\ndescription: Use this agent when [condition]. Typical triggers include [trigger 1] and [trigger 2]. See \"When to invoke\" in the agent body.\nmodel: inherit\ncolor: blue\n---\n\nYou are an agent that [does X].\n\n## When to invoke\n\n- **[Scenario A].** [Description.]\n- **[Scenario B].** [Description.]\n\nProcess:\n1. [Step 1]\n2. [Step 2]\n\nOutput: [What to provide]\n```\n\n### Frontmatter Fields Summary\n\n| Field | Required | Format | Example |\n|-------|----------|--------|---------|\n| name | Yes | lowercase-hyphens | code-reviewer |\n| description | Yes | Prose triggers | Use when... Typical triggers include... |\n| model | Yes | inherit\u002Fsonnet\u002Fopus\u002Fhaiku | inherit |\n| color | Yes | Color name | blue |\n| tools | No | Array of tool names | [\"Read\", \"Grep\"] |\n\n### Best Practices\n\n**DO:**\n- ✅ Name 2-4 trigger scenarios in the description (as prose)\n- ✅ Put detailed worked scenarios in a \"When to invoke\" body section, as prose bullets\n- ✅ Write specific triggering conditions\n- ✅ Use `inherit` for model unless specific need\n- ✅ Choose appropriate tools (least privilege)\n- ✅ Write clear, structured system prompts\n- ✅ Test agent triggering thoroughly\n\n**DON'T:**\n- ❌ Use generic descriptions without trigger scenarios\n- ❌ Omit triggering conditions\n- ❌ Give all agents same color\n- ❌ Grant unnecessary tool access\n- ❌ Write vague system prompts\n- ❌ Skip testing\n\n## Additional Resources\n\n### Reference Files\n\nFor detailed guidance, consult:\n\n- **`references\u002Fsystem-prompt-design.md`** - Complete system prompt patterns\n- **`references\u002Ftriggering-examples.md`** - Example formats and best practices\n- **`references\u002Fagent-creation-system-prompt.md`** - The exact prompt from Claude Code\n\n### Example Files\n\nWorking examples in `examples\u002F`:\n\n- **`agent-creation-prompt.md`** - AI-assisted agent generation template\n- **`complete-agent-examples.md`** - Full agent examples for different use cases\n\n### Utility Scripts\n\nDevelopment tools in `scripts\u002F`:\n\n- **`validate-agent.sh`** - Validate agent file structure\n- **`test-agent-trigger.sh`** - Test if agent triggers correctly\n\n## Implementation Workflow\n\nTo create an agent for a plugin:\n\n1. Define agent purpose and triggering conditions\n2. Choose creation method (AI-assisted or manual)\n3. Create `agents\u002Fagent-name.md` file\n4. Write frontmatter with all required fields\n5. Write system prompt following best practices\n6. Name 2-4 trigger scenarios in description (prose) and detail them in a \"When to invoke\" body section\n7. Validate with `scripts\u002Fvalidate-agent.sh`\n8. Test triggering with real scenarios\n9. Document agent in plugin README\n\nFocus on clear triggering conditions and comprehensive system prompts for autonomous operation.\n","# Agent Development for Claude Code Plugins\n\n## Overview\n\nAgents are autonomous subprocesses that handle complex, multi-step tasks independently. Understanding agent structure, triggering conditions, and system prompt design enables creating powerful autonomous capabilities.\n\n**Key concepts:**\n- Agents are FOR autonomous work, commands are FOR user-initiated actions\n- Markdown file format with YAML frontmatter\n- Triggering via description field with examples\n- System prompt defines agent behavior\n- Model and color customization\n\n## Agent File Structure\n\n### Complete Format\n\n```markdown\n---\nname: agent-identifier\ndescription: Use this agent when [triggering conditions]. Typical triggers include [scenario 1 in prose], [scenario 2 in prose], and [scenario 3 in prose]. See \"When to invoke\" in the agent body for worked scenarios.\nmodel: inherit\ncolor: blue\ntools: [\"Read\", \"Write\", \"Grep\"]\n---\n\nYou are [agent role description]...\n\n## When to invoke\n\n[Two to four representative scenarios written as prose, e.g.:]\n- **[Scenario name].** [What the situation looks like and what the agent should do.]\n- **[Scenario name].** [Same.]\n\n**Your Core Responsibilities:**\n1. [Responsibility 1]\n2. [Responsibility 2]\n\n**Analysis Process:**\n[Step-by-step workflow]\n\n**Output Format:**\n[What to return]\n```\n\n## Frontmatter Fields\n\n### name (required)\n\nAgent identifier used for namespacing and invocation.\n\n**Format:** lowercase, numbers, hyphens only\n**Length:** 3-50 characters\n**Pattern:** Must start and end with alphanumeric\n\n**Good examples:**\n- `code-reviewer`\n- `test-generator`\n- `api-docs-writer`\n- `security-analyzer`\n\n**Bad examples:**\n- `helper` (too generic)\n- `-agent-` (starts\u002Fends with hyphen)\n- `my_agent` (underscores not allowed)\n- `ag` (too short, \u003C 3 chars)\n\n### description (required)\n\nDefines when Claude should trigger this agent. **This is the most critical field** — it is loaded into context whenever the agent is registered, so the harness can decide when to dispatch.\n\n**Must include:**\n1. Triggering conditions (\"Use this agent when...\")\n2. A short prose summary of the typical trigger scenarios\n3. A pointer to a \"When to invoke\" section in the agent body for the detailed worked scenarios\n\n**Format:**\n```\nUse this agent when [conditions]. Typical triggers include [scenario 1 in prose], [scenario 2 in prose], and [scenario 3 in prose]. See \"When to invoke\" in the agent body for worked scenarios.\n```\n\n**Best practices:**\n- Name 2-4 trigger scenarios in the prose summary\n- Cover both proactive (assistant invokes itself) and reactive (user requests) triggering\n- Cover different phrasings of the same intent\n- Be specific about when NOT to use the agent\n- Put detailed scenarios in the body under \"When to invoke\" as a bullet list of prose descriptions\n\n### model (required)\n\nWhich model the agent should use.\n\n**Options:**\n- `inherit` - Use same model as parent (recommended)\n- `sonnet` - Claude Sonnet (balanced)\n- `opus` - Claude Opus (most capable, expensive)\n- `haiku` - Claude Haiku (fast, cheap)\n\n**Recommendation:** Use `inherit` unless agent needs specific model capabilities.\n\n### color (required)\n\nVisual identifier for agent in UI.\n\n**Options:** `blue`, `cyan`, `green`, `yellow`, `magenta`, `red`\n\n**Guidelines:**\n- Choose distinct colors for different agents in same plugin\n- Use consistent colors for similar agent types\n- Blue\u002Fcyan: Analysis, review\n- Green: Success-oriented tasks\n- Yellow: Caution, validation\n- Red: Critical, security\n- Magenta: Creative, generation\n\n### tools (optional)\n\nRestrict agent to specific tools.\n\n**Format:** Array of tool names\n\n```yaml\ntools: [\"Read\", \"Write\", \"Grep\", \"Bash\"]\n```\n\n**Default:** If omitted, agent has access to all tools\n\n**Best practice:** Limit tools to minimum needed (principle of least privilege)\n\n**Common tool sets:**\n- Read-only analysis: `[\"Read\", \"Grep\", \"Glob\"]`\n- Code generation: `[\"Read\", \"Write\", \"Grep\"]`\n- Testing: `[\"Read\", \"Bash\", \"Grep\"]`\n- Full access: Omit field or use `[\"*\"]`\n\n## System Prompt Design\n\nThe markdown body becomes the agent's system prompt. Write in second person, addressing the agent directly.\n\n### Structure\n\n**Standard template:**\n```markdown\nYou are [role] specializing in [domain].\n\n**Your Core Responsibilities:**\n1. [Primary responsibility]\n2. [Secondary responsibility]\n3. [Additional responsibilities...]\n\n**Analysis Process:**\n1. [Step one]\n2. [Step two]\n3. [Step three]\n[...]\n\n**Quality Standards:**\n- [Standard 1]\n- [Standard 2]\n\n**Output Format:**\nProvide results in this format:\n- [What to include]\n- [How to structure]\n\n**Edge Cases:**\nHandle these situations:\n- [Edge case 1]: [How to handle]\n- [Edge case 2]: [How to handle]\n```\n\n### Best Practices\n\n✅ **DO:**\n- Write in second person (\"You are...\", \"You will...\")\n- Be specific about responsibilities\n- Provide step-by-step process\n- Define output format\n- Include quality standards\n- Address edge cases\n- Keep under 10,000 characters\n\n❌ **DON'T:**\n- Write in first person (\"I am...\", \"I will...\")\n- Be vague or generic\n- Omit process steps\n- Leave output format undefined\n- Skip quality guidance\n- Ignore error cases\n\n## Creating Agents\n\n### Method 1: AI-Assisted Generation\n\nUse this prompt pattern (extracted from Claude Code):\n\n```\nCreate an agent configuration based on this request: \"[YOUR DESCRIPTION]\"\n\nRequirements:\n1. Extract core intent and responsibilities\n2. Design expert persona for the domain\n3. Create comprehensive system prompt with:\n   - Clear behavioral boundaries\n   - Specific methodologies\n   - Edge case handling\n   - Output format\n   - A \"When to invoke\" section listing 2-4 trigger scenarios as prose bullets\n4. Create identifier (lowercase, hyphens, 3-50 chars)\n5. Write description with triggering conditions and a short prose summary of trigger scenarios\n\nReturn JSON with:\n{\n  \"identifier\": \"agent-name\",\n  \"whenToUse\": \"Use this agent when... Typical triggers include [...]. See \\\"When to invoke\\\" in the agent body.\",\n  \"systemPrompt\": \"You are...\"\n}\n```\n\nThen convert to agent file format with frontmatter.\n\nSee `examples\u002Fagent-creation-prompt.md` for complete template.\n\n### Method 2: Manual Creation\n\n1. Choose agent identifier (3-50 chars, lowercase, hyphens)\n2. Write description with examples\n3. Select model (usually `inherit`)\n4. Choose color for visual identification\n5. Define tools (if restricting access)\n6. Write system prompt with structure above\n7. Save as `agents\u002Fagent-name.md`\n\n## Validation Rules\n\n### Identifier Validation\n\n```\n✅ Valid: code-reviewer, test-gen, api-analyzer-v2\n❌ Invalid: ag (too short), -start (starts with hyphen), my_agent (underscore)\n```\n\n**Rules:**\n- 3-50 characters\n- Lowercase letters, numbers, hyphens only\n- Must start and end with alphanumeric\n- No underscores, spaces, or special characters\n\n### Description Validation\n\n**Length:** 10-5,000 characters\n**Must include:** Triggering conditions and examples\n**Best:** 200-1,000 characters with 2-4 examples\n\n### System Prompt Validation\n\n**Length:** 20-10,000 characters\n**Best:** 500-3,000 characters\n**Structure:** Clear responsibilities, process, output format\n\n## Agent Organization\n\n### Plugin Agents Directory\n\n```\nplugin-name\u002F\n└── agents\u002F\n    ├── analyzer.md\n    ├── reviewer.md\n    └── generator.md\n```\n\nAll `.md` files in `agents\u002F` are auto-discovered.\n\n### Namespacing\n\nAgents are namespaced automatically:\n- Single plugin: `agent-name`\n- With subdirectories: `plugin:subdir:agent-name`\n\n## Testing Agents\n\n### Test Triggering\n\nCreate test scenarios to verify agent triggers correctly:\n\n1. Write agent with specific triggering examples\n2. Use similar phrasing to examples in test\n3. Check Claude loads the agent\n4. Verify agent provides expected functionality\n\n### Test System Prompt\n\nEnsure system prompt is complete:\n\n1. Give agent typical task\n2. Check it follows process steps\n3. Verify output format is correct\n4. Test edge cases mentioned in prompt\n5. Confirm quality standards are met\n\n## Quick Reference\n\n### Minimal Agent\n\n```markdown\n---\nname: simple-agent\ndescription: Use this agent when [condition]. Typical triggers include [trigger 1] and [trigger 2]. See \"When to invoke\" in the agent body.\nmodel: inherit\ncolor: blue\n---\n\nYou are an agent that [does X].\n\n## When to invoke\n\n- **[Scenario A].** [Description.]\n- **[Scenario B].** [Description.]\n\nProcess:\n1. [Step 1]\n2. [Step 2]\n\nOutput: [What to provide]\n```\n\n### Frontmatter Fields Summary\n\n| Field | Required | Format | Example |\n|-------|----------|--------|---------|\n| name | Yes | lowercase-hyphens | code-reviewer |\n| description | Yes | Prose triggers | Use when... Typical triggers include... |\n| model | Yes | inherit\u002Fsonnet\u002Fopus\u002Fhaiku | inherit |\n| color | Yes | Color name | blue |\n| tools | No | Array of tool names | [\"Read\", \"Grep\"] |\n\n### Best Practices\n\n**DO:**\n- ✅ Name 2-4 trigger scenarios in the description (as prose)\n- ✅ Put detailed worked scenarios in a \"When to invoke\" body section, as prose bullets\n- ✅ Write specific triggering conditions\n- ✅ Use `inherit` for model unless specific need\n- ✅ Choose appropriate tools (least privilege)\n- ✅ Write clear, structured system prompts\n- ✅ Test agent triggering thoroughly\n\n**DON'T:**\n- ❌ Use generic descriptions without trigger scenarios\n- ❌ Omit triggering conditions\n- ❌ Give all agents same color\n- ❌ Grant unnecessary tool access\n- ❌ Write vague system prompts\n- ❌ Skip testing\n\n## Additional Resources\n\n### Reference Files\n\nFor detailed guidance, consult:\n\n- **`references\u002Fsystem-prompt-design.md`** - Complete system prompt patterns\n- **`references\u002Ftriggering-examples.md`** - Example formats and best practices\n- **`references\u002Fagent-creation-system-prompt.md`** - The exact prompt from Claude Code\n\n### Example Files\n\nWorking examples in `examples\u002F`:\n\n- **`agent-creation-prompt.md`** - AI-assisted agent generation template\n- **`complete-agent-examples.md`** - Full agent examples for different use cases\n\n### Utility Scripts\n\nDevelopment tools in `scripts\u002F`:\n\n- **`validate-agent.sh`** - Validate agent file structure\n- **`test-agent-trigger.sh`** - Test if agent triggers correctly\n\n## Implementation Workflow\n\nTo create an agent for a plugin:\n\n1. Define agent purpose and triggering conditions\n2. Choose creation method (AI-assisted or manual)\n3. Create `agents\u002Fagent-name.md` file\n4. Write frontmatter with all required fields\n5. Write system prompt following best practices\n6. Name 2-4 trigger scenarios in description (prose) and detail them in a \"When to invoke\" body section\n7. Validate with `scripts\u002Fvalidate-agent.sh`\n8. Test triggering with real scenarios\n9. Document agent in plugin README\n\nFocus on clear triggering conditions and comprehensive system prompts for autonomous operation.\n",{"name":5,"version":49,"description":7},"0.1.0","main","dceb15d50832e6adf380bedc26dde03d251324b9","plugins\u002Fplugin-dev\u002Fskills\u002Fagent-development",4875,"https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fclaude-plugins-official","https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fplugins","https:\u002F\u002Fgithub.com\u002Fanthropics","https:\u002F\u002Fanthropic.com","2026-11-12 03:40:35",{"source":8,"skillName":5,"commands":60,"prompt":70,"zipAvailable":71},[61,64,67],{"tool":62,"command":63},"npx","npx skills add anthropics\u002Fclaude-plugins-official --skill agent-development",{"tool":65,"command":66},"bunx","bunx skills add anthropics\u002Fclaude-plugins-official --skill agent-development",{"tool":68,"command":69},"pnpm","pnpm dlx skills add anthropics\u002Fclaude-plugins-official --skill agent-development","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fclaude-plugins-official\u002Fblob\u002Fmain\u002Fplugins\u002Fplugin-dev\u002Fskills\u002Fagent-development\u002FSKILL.md",true,[73,79,83,87,91,95],{"path":74,"resourceType":75,"resourceTypeDesc":76,"sizeBytes":77,"sourceHtmlUrl":78},"examples\u002Fagent-creation-prompt.md","REFERENCE","参考文档",9838,"https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fclaude-plugins-official\u002Fblob\u002Fmain\u002Fplugins\u002Fplugin-dev\u002Fskills\u002Fagent-development\u002Fexamples\u002Fagent-creation-prompt.md",{"path":80,"resourceType":75,"resourceTypeDesc":76,"sizeBytes":81,"sourceHtmlUrl":82},"examples\u002Fcomplete-agent-examples.md",13916,"https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fclaude-plugins-official\u002Fblob\u002Fmain\u002Fplugins\u002Fplugin-dev\u002Fskills\u002Fagent-development\u002Fexamples\u002Fcomplete-agent-examples.md",{"path":84,"resourceType":75,"resourceTypeDesc":76,"sizeBytes":85,"sourceHtmlUrl":86},"references\u002Fagent-creation-system-prompt.md",9749,"https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fclaude-plugins-official\u002Fblob\u002Fmain\u002Fplugins\u002Fplugin-dev\u002Fskills\u002Fagent-development\u002Freferences\u002Fagent-creation-system-prompt.md",{"path":88,"resourceType":75,"resourceTypeDesc":76,"sizeBytes":89,"sourceHtmlUrl":90},"references\u002Fsystem-prompt-design.md",9998,"https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fclaude-plugins-official\u002Fblob\u002Fmain\u002Fplugins\u002Fplugin-dev\u002Fskills\u002Fagent-development\u002Freferences\u002Fsystem-prompt-design.md",{"path":92,"resourceType":75,"resourceTypeDesc":76,"sizeBytes":93,"sourceHtmlUrl":94},"references\u002Ftriggering-examples.md",8868,"https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fclaude-plugins-official\u002Fblob\u002Fmain\u002Fplugins\u002Fplugin-dev\u002Fskills\u002Fagent-development\u002Freferences\u002Ftriggering-examples.md",{"path":96,"resourceType":97,"resourceTypeDesc":98,"sizeBytes":99,"sourceHtmlUrl":100},"scripts\u002Fvalidate-agent.sh","SCRIPT","脚本",5796,"https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fclaude-plugins-official\u002Fblob\u002Fmain\u002Fplugins\u002Fplugin-dev\u002Fskills\u002Fagent-development\u002Fscripts\u002Fvalidate-agent.sh",[102,127,152,176,201,217],{"skillNo":103,"name":104,"nameZh":6,"description":105,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":106,"sourceHtmlUrl":107,"categories":108,"keywords":110,"publishedAt":44},"01M718BVSPNAK6S5NZCYFDBG6E","skill-creator","Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.","https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fclaude-plugins-official\u002Fblob\u002Fmain\u002Fplugins\u002Fskill-creator\u002Fskills\u002Fskill-creator\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fclaude-plugins-official\u002Fblob\u002Fmain\u002Fplugins\u002Fskill-creator\u002Fskills\u002Fskill-creator\u002FSKILL.md",[109],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},[111,112,113,114,115,116,42,117,39,118,119,120,121,122,123,124,32,125,126,30],"skill creator","creator","performance","improve","run evals","new","analysis","optimize","variance analysis","run","edit","benchmark","existing skill","evals","accuracy","scratch",{"skillNo":128,"name":129,"nameZh":6,"description":130,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":131,"categories":132,"keywords":138,"publishedAt":44},"01M718BSVAWJRFN6MRCZZ0QX4X","session-report","Generate an explorable HTML report of Claude Code session usage (tokens, cache, subagents, skills, expensive prompts) from ~\u002F.claude\u002Fprojects transcripts.","https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fclaude-plugins-official\u002Fblob\u002Fmain\u002Fplugins\u002Fsession-report\u002Fskills\u002Fsession-report\u002FSKILL.md",[133],{"categoryNo":134,"categoryName":135,"taxonomyType":136,"taxonomyTypeDesc":137},"01JBSEEDCAT000000000000034","行政与办公人员","PROFESSION","适用职业",[139,140,141,142,143,144,145,146,147,148,149,41,150,151],"session","report","session report","projects","usage","generate","tokens","html report","transcripts","prompts","cache","html","subagents",{"skillNo":153,"name":154,"nameZh":6,"description":155,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":156,"categories":157,"keywords":161,"publishedAt":44},"01M718BPKHVD214FAS1T4E5XCE","receipts","Generate a personal Claude Code usage & impact report (\"receipts\") from this machine's local session transcripts — for justifying Claude Code usage\u002Fspend to a manager, self-review, or \"what have I been using this for\" check-ins. Mines ~\u002F.claude\u002Fprojects locally (no extra API calls beyond one final write-up), cross-references local git history, and writes a markdown report plus a self-contained HTML receipt to your home directory. Use when the user asks for \"receipts\", an \"impact report\", \"usage report\", wants to \"show my Claude Code activity\", \"prove the value of Claude Code\", or runs `\u002Freceipts`.","https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fclaude-plugins-official\u002Fblob\u002Fmain\u002Fplugins\u002Freceipts\u002Fskills\u002Freceipts\u002FSKILL.md",[158],{"categoryNo":159,"categoryName":160,"taxonomyType":136,"taxonomyTypeDesc":137},"01JBSEEDCAT000000000000030","开发工程师",[154,142,162,163,164,165,42,41,166,167,168,169,144,170,171,37,172,173,174,175],"show","up","local","impact","api","review","self","references","markdown","directory","git","activity","my","manager",{"skillNo":177,"name":178,"nameZh":6,"description":179,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":180,"categories":181,"keywords":182,"publishedAt":44},"01M718BJMX0T4691FJQ62GYY5C","project-artifact","Generate and publish a project status artifact — an opinionated, tabbed status page for a project too big for one update (overview & success criteria, the workstream sequence, next steps, plus background, plan, risks & open questions, and decisions\u002FFAQ when they earn a tab) — published with the built-in Artifact tool to a default-private claude.ai page the user can share with teammates. Use when a piece of work spans several workstreams and you want a shareable overview kept current. Each artifact is backed by a small per-project config in the plugin data dir, so refreshing it re-gathers live state, redeploys the same URL, and reports only the delta. For software projects whose workstreams are PRs, also read swe.md (the X.Y PR-numbering convention; pulling PR state with gh\u002Fgit; a per-PR detail block). Needs the built-in Artifact tool (claude.ai login). Not for single-PR changes or public docs.","https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fclaude-plugins-official\u002Fblob\u002Fmain\u002Fplugins\u002Fproject-artifact\u002Fskills\u002Fproject-artifact\u002FSKILL.md",[],[183,184,185,186,142,187,188,189,190,191,192,42,193,194,195,196,197,198,199,200],"project","artifact","project artifact","project status","re","default","next steps","open","only","SWE","URL","you","reports","publish","work","plan","status page","next",{"skillNo":202,"name":203,"nameZh":6,"description":204,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":205,"categories":206,"keywords":208,"publishedAt":44},"01M718BF1TTFP5CM0S5HBW64FJ","skill-development","This skill should be used when the user wants to \"create a skill\", \"add a skill to plugin\", \"write a new skill\", \"improve skill description\", \"organize skill content\", or needs guidance on skill structure, progressive disclosure, or skill development best practices for Claude Code plugins.","https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fclaude-plugins-official\u002Fblob\u002Fmain\u002Fplugins\u002Fplugin-dev\u002Fskills\u002Fskill-development\u002FSKILL.md",[207],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},[24,209,210,114,116,211,42,41,39,38,37,212,213,35,33,32,214,29,215,216],"skill development","skill to plugin","improve skill","organize","new skill","progressive","progressive disclosure","plugin",{"skillNo":218,"name":219,"nameZh":6,"description":220,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":221,"categories":222,"keywords":224,"publishedAt":44},"01M718BCARZ2J4PW0X3AJCZ8MZ","plugin-structure","This skill should be used when the user asks to \"create a plugin\", \"scaffold a plugin\", \"understand plugin structure\", \"organize plugin components\", \"set up plugin.json\", \"use ${CLAUDE_PLUGIN_ROOT}\", \"add commands\u002Fagents\u002Fskills\u002Fhooks\", \"configure auto-discovery\", or needs guidance on plugin directory layout, manifest configuration, component organization, file naming conventions, or Claude Code plugin architecture best practices.","https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fclaude-plugins-official\u002Fblob\u002Fmain\u002Fplugins\u002Fplugin-dev\u002Fskills\u002Fplugin-structure\u002FSKILL.md",[223],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},[38,225,216,226,227,228,163,229,230,42,231,41,232,233,39,234,235,236,171,237],"Plugin Structure","SET","plugin architecture","hooks","auto-discovery","auto","configure","components","architecture","component","layout","organization","commands"]