Package Exports
- @agentpowers/mcp-server
- @agentpowers/mcp-server/dist/index.js
This package does not declare an exports field, so the exports above have been automatically detected and optimized by JSPM instead. If any package subpath is missing, it is recommended to post an issue to the original package (@agentpowers/mcp-server) to support the "exports" field. If that is not possible, create a JSPM override to customize the exports field for this package.
Readme
@agentpowers/mcp-server
MCP server for the AgentPowers marketplace. Search, browse, install, and manage paid skills and agents directly inside Claude Code, Claude Cowork, and claude.ai conversations.
Tools
| Tool | Description |
|---|---|
search_marketplace |
Search for skills and agents by query, category, or type |
get_skill_details |
Get detailed info (price, security status, platforms, versions) for a skill |
install_skill |
Download and install a skill after security check |
check_purchase_status |
Check payment status and retrieve license code |
check_installed |
List all installed skills with version and edit status |
uninstall_skill |
Remove an installed skill and its version pin |
check_for_updates |
Check installed skills for newer versions |
Usage with Claude Code
Install the AgentPowers plugin, which includes this MCP server:
claude /plugin install agentpowers
claude /plugin enable agentpowersThe MCP server runs automatically via stdio transport when the plugin is active.
Usage as standalone MCP server
npx @agentpowers/mcp-serverOr install globally:
npm install -g @agentpowers/mcp-server
agentpowers-mcpAuthentication
Most tools work without authentication. Installing paid skills and checking purchase status require a CLI auth token stored at ~/.agentpowers/auth.json. Authenticate via:
ap loginConfiguration
The server connects to https://api.agentpowers.ai by default. Override with the AGENTPOWERS_API_URL environment variable for development.
Security
Every skill passes an 8-layer security pipeline before listing: file type validation, regex pattern scanning, ClamAV, Snyk, GuardDog, VirusTotal, AI security review, and agent-specific checks. Results are cached by content hash and displayed in skill details.
License
MIT