Forage MCP Server

MCPOpen SourceMIT31.6

by isaac-levine • Uncategorized

An MCP server that enables AI agents to discover, install, and learn to use new tools automatically without restarts or manual configuration.

Example Use Cases

1
Dynamically discover and install new tools without manual intervention.
2
Expand their capabilities on the fly by adding MCP servers for new tasks.
3
Persistent knowledge of installed tools across sessions.

Description

Forage acts as a gateway MCP server that allows AI agents to dynamically find and install new MCP servers and tools on demand. It searches multiple registries, installs tools as subprocesses, and immediately exposes their capabilities to the agent. This self-improving mechanism helps agents overcome limitations by automatically expanding their toolset and persisting knowledge for future sessions.

Capabilities(6 total)

Tools (6)

forage_search

Search for MCP servers and tools across registries (Official MCP Registry, Smithery, npm). Describe what capability you need and get ranked results.

query:string*sources:array

forage_evaluate

Get detailed information about a specific package — version, downloads, README summary, install command. Use this before installing to verify quality.

packageName:string*

forage_install

Install an MCP server and start it as a proxied subprocess. The server's tools become available immediately via Forage — no restart needed. IMPORTANT: Set confirm=true to proceed with installation (user will be prompted to approve the tool call).

packageName:string*command:stringargs:arrayenv:objectconfirm:boolean*

forage_learn

Write usage instructions for an installed tool to agent rule files (CLAUDE.md, AGENTS.md, .cursor/rules/). This persists knowledge across sessions so the agent remembers how to use the tool next time.

toolName:string*rules:string*target:string

forage_status

List all tools that Forage has installed and is currently proxying. Shows which servers are running, what tools they provide, and their install status.

Quick Actions

View on GitHub

Security

Scanned 2 month(s) ago

Risk Level

MINIMAL

Read-only data retrieval, no side effects

Trust Score

C57/100

10/17 checks passed

Scores are informational only and provided “as is” without warranty. AgentHotspot assumes no liability for actions taken based on these ratings.

Quick Stats

Service TypeMCP

Pricing ModelFree

Capabilities6 Tools / 0 Prompts / 0 Resources

Ownerisaac-levine

CategoryUncategorized

DependenciesStandalone

Set Your Username