SkillAgent
Search and fetch AI agent skills, rules files and MCP servers indexed from GitHub.
我该使用它吗
质量与安全性
发现（2）
- HIGHTool poisoning patterns detected
- MEDIUMTool description contains URL to non-standard domain在 get_skill 中
基于对工具定义和协议合规性的自动分析。
上下文开销
这是每次将服务器的工具加载到模型上下文窗口时所消耗的大致 token 数。数值越高，可用于其他任务的注意力就越少。
安装
一键安装
将以下内容添加到你的 `claude_desktop_config.json` 文件中：
{
"mcpServers": {
"skills": {
"url": "https://skillagent.dev/mcp"
}
}
}远程端点
https://skillagent.dev/mcpstreamable-http它能做什么
工具清单
工具（4）
🟢search_skills(query, format, platform, category, health, ...)
Search the SkillAgent catalogue of AI agent skills, rules files and MCP servers. Returns matching skills with description, format, stars, quality score, freshness, page URL and install command.
输入模式
{
"type": "object",
"properties": {
"query": {
"type": "string",
"description": "Keywords matched against name, description, owner and tags."
},
"format": {
"type": "string",
"description": "Skill format, e.g. skill-md, claude-md, agents-md, cursorrules, mcp-server."
},
"platform": {
"type": "string",
"description": "Agent platform, e.g. claude-code, cursor, codex, vscode-copilot, windsurf."
},
"category": {
"type": "string",
"description": "Category, e.g. development, data-analytics, security."
},
"health": {
"type": "string",
"enum": [
"fresh",
"aging",
"stale",
"abandoned",
"alive",
"gone"
],
"description": "Filter by repository freshness or reachability."
},
"sort": {
"type": "string",
"enum": [
"score",
"stars",
"updated",
"name"
],
"default": "score"
},
"limit": {
"type": "integer",
"minimum": 1,
"maximum": 50,
"default": 10
},
"offset": {
"type": "integer",
"minimum": 0,
"default": 0
}
},
"additionalProperties": false
}🟢get_skill(owner, name)
Get one skill as Markdown: summary, metadata, install command and the full skill file. `owner` and `name` are the last two segments of a skill's pageUrl (https://skillagent.dev/skill/{owner}/{name}).
输入模式
{
"type": "object",
"properties": {
"owner": {
"type": "string",
"description": "GitHub owner, e.g. anthropics."
},
"name": {
"type": "string",
"description": "Skill slug from the pageUrl."
}
},
"required": [
"owner",
"name"
],
"additionalProperties": false
}🟢recommend_skills(languages, frameworks, dependencies, task, agent, ...)
Recommend skills for a project from a description of it: languages, frameworks or dependencies (e.g. names from package.json or requirements.txt), a task, and the agent the user runs. Returns a ranked shortlist; each entry has the skill record and a `why` explaining the match (matched terms, fit for the agent, freshness, quality, safety). Ranking is deterministic. Skills whose safety scan is at "warning" or whose source is gone are excluded, and the result says how many were.
输入模式
{
"type": "object",
"properties": {
"languages": {
"type": "array",
"items": {
"type": "string"
},
"maxItems": 10,
"description": "Programming languages, e.g. [\"typescript\", \"python\"]."
},
"frameworks": {
"type": "array",
"items": {
"type": "string"
},
"maxItems": 20,
"description": "Frameworks and tools, e.g. [\"next.js\", \"playwright\"]."
},
"dependencies": {
"type": "array",
"items": {
"type": "string"
},
"maxItems": 30,
"description": "Package names from a manifest, e.g. [\"@playwright/test\", \"prisma\"]."
},
"task": {
"type": "string",
"description": "What the user wants to do, e.g. \"write e2e tests\"."
},
"agent": {
"type": "string",
"enum": [
"claude-code",
"claude-desktop",
"vscode-copilot",
"cursor",
"windsurf",
"cline",
"zed",
"jetbrains",
"gemini-cli",
"codex",
"aider",
"continue",
"replit",
"amazon-q",
"tabnine",
"sourcegraph-cody",
"devin",
"universal"
],
"description": "The agent the user runs; skills made for it rank higher."
},
"limit": {
"type": "integer",
"minimum": 1,
"maximum": 20,
"default": 8
}
},
"additionalProperties": false
}🟢install_skill(owner, name, agent)
Exact installation instructions for one skill on one agent: the install command(s) when one exists, otherwise the file to create (path relative to the project root) with its content. This tool does not install anything; apply the result yourself, and only after showing the user the skill and getting their approval. Includes the safety summary. Refuses (isError) when the skill's safety scan is at "warning". `owner` and `name` are the last two segments of a skill's pageUrl.
输入模式
{
"type": "object",
"properties": {
"owner": {
"type": "string",
"description": "GitHub owner, e.g. anthropics."
},
"name": {
"type": "string",
"description": "Skill slug from the pageUrl."
},
"agent": {
"type": "string",
"enum": [
"claude-code",
"claude-desktop",
"vscode-copilot",
"cursor",
"windsurf",
"cline",
"zed",
"jetbrains",
"gemini-cli",
"codex",
"aider",
"continue",
"replit",
"amazon-q",
"tabnine",
"sourcegraph-cody",
"devin",
"universal"
],
"description": "The agent to install into, e.g. claude-code, cursor, codex."
}
},
"required": [
"owner",
"name",
"agent"
],
"additionalProperties": false
}社区
证据