@alfe.ai/mcp-tools
v0.2.14
Published
Shared MCP tool implementations for Alfe's local CLI-bundled MCP server and (later) the Fly-deployed services/mcp. All tools use AgentApiClient against public /agent/... endpoints — never InternalClient.
Readme
@alfe.ai/mcp-tools
Shared MCP tool implementations for Alfe's local CLI-bundled MCP server. All
Alfe API calls use AgentApiClient against public agent endpoints — never
InternalClient or raw authenticated fetch.
The package exports registrars for integrations, memory, one-shot voice, and outbound SMS. The server chooses which registrars to expose for each runtime profile.
Model-controlled file arguments are workspace-relative and realpath-contained. Memory and audio inputs are size-bounded; voice output creates a new file and never overwrites an existing path. Tool schemas mirror the downstream public API's string, enum, and numeric limits.
Part of Alfe — the operating system for AI agents: build, deploy, and run agents with persistent memory, identity, integrations, and channels. See the documentation to get started.
Install
npm install @alfe.ai/mcp-toolsSee DEVELOPING.md for the tool and filesystem contracts.
Links
- 🌐 Website: https://alfe.ai
- 📚 Docs: https://docs.alfe.ai
