菜单
Claude Code and Anthropic ecosystem expert. Provides official patterns for creating agents, skills, hooks, and commands. Use when creating or validating Claude Code artifacts, asking about Claude Code features, MCP integration, or tool usage patterns.
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。
Expert at understanding and working with Hermes Agent — its memory system, skills, cron jobs, tools, and behavioral conventions. Use when configuring Hermes, writing skills for it, debugging its behavior, or understanding what it can and cannot do.
Set up, run, tune, and debug pxpipe (pxpipe-proxy) — a local proxy that cuts Claude Code token costs ~59–70% by rendering bulky context (tool results, old history, system prompt) as compressed PNG images before requests leave the machine. Use whenever the user mentions pxpipe, pxpipe-proxy, imaging context, PNG context compression, reducing Claude Code token costs/bill via a proxy, ANTHROPIC_BASE_URL proxies for token savings, or asks to verify pxpipe savings or troubleshoot misread/confabulated values while running it.
Expert on the 1Password CLI (the `op` command) for managing 1Password from the terminal and securely loading secrets into scripts and CI. Use when the user wants to read/inject secrets, use secret references (op://vault/item/field), run `op run` or `op inject` or `op read`, manage items/vaults, set up service accounts for automation, configure shell plugins (`op plugin`), use the 1Password SSH agent or git commit signing, or enable biometric/desktop-app integration. Triggers on mentions of 1Password CLI, `op` command, op cli, secret references, op run, op inject, op read, service account, secrets automation, 1Password vault/item from terminal, op plugin, 1Password SSH agent.
Expert on browser automation with AI agents using Vercel's agent-browser CLI. Use when the user wants to automate browsers, scrape websites, capture screenshots, or build AI-powered web workflows. Triggers on mentions of agent-browser, browser automation, headless browsing, web scraping with AI.
Expert at Agent Skills - both the open agentskills.io format and Claude's official extensions (Claude Code, Agent SDK, Claude API). Use when creating SKILL.md files, understanding the specification, validating skills, choosing frontmatter fields, or integrating skills into agents. Triggers on mentions of agent skills, SKILL.md, skill format, slash commands, skill frontmatter, agent extensibility.
Expert on Vercel's AI SDK for building AI-powered applications. Use when user wants to integrate LLMs, stream text, generate structured data, use tools, or build agents. Triggers on mentions of ai-sdk, vercel ai, generateText, streamText, useChat, tool calling.
| name | claude-code-expert |
| description | Claude Code and Anthropic ecosystem expert. Provides official patterns for creating agents, skills, hooks, and commands. Use when creating or validating Claude Code artifacts, asking about Claude Code features, MCP integration, or tool usage patterns. |
| allowed-tools | Read, Grep, Glob, WebFetch |
Provides authoritative knowledge for Claude Code development, ensuring all artifacts follow official best practices.
docs/patterns/agent-creation.md - Agent creation guidedocs/patterns/skill-creation.md - Skill creation guidedocs/patterns/hook-creation.md - Hook creation guidedocs/patterns/hook-advanced.md - Advanced hook patterns (v2.1.0+)docs/patterns/command-creation.md - Command creation guidedocs/features/tool-usage.md - Read, Write, Edit, Grep, Glob, Bashdocs/features/mcp-integration.md - MCP servers and integrationdocs/features/code-execution.md - Code execution with MCPdocs/validation/agent-checklist.md - Agent quality checksdocs/validation/skill-checklist.md - Skill quality checksdocs/validation/hook-checklist.md - Hook quality checksdocs/validation/command-checklist.md - Command quality checksdocs/ecosystem/claude-versions.md - Model versionsdocs/ecosystem/model-capabilities.md - Capability comparisonUser: "Create an agent for code security reviews"
→ Load docs/patterns/agent-creation.md
→ Provide structure, frontmatter, system prompt guidance
→ After creation, validate against docs/validation/agent-checklist.md
User: "Validate my authentication hook"
→ Load docs/validation/hook-checklist.md
→ Review against all checklist items
→ Report issues with recommendations
External documentation is cached from official sources. To update:
# Check freshness
./scripts/check-updates.sh --verbose
# Sync stale sources
./scripts/sync-sources.sh
# Force refresh
./scripts/sync-sources.sh --force
Sources defined in sources/registry.json: