kumo-rfm-mcp
Verified Safeby kumo-ai
Overview
This server empowers AI assistants with KumoRFM intelligence by providing tools to build, manage, and visualize relational graphs, convert natural language to PQL queries, and execute/evaluate/explain predictions from the KumoRFM (Relational Foundation Model) for tasks like missing value imputation and temporal forecasting.
Installation
python -m kumo_rfm_mcp.serverEnvironment Variables
- KUMO_API_KEY
Security Notes
The server uses `os.getenv` for API keys, which is a secure practice. It utilizes `subprocess.check_call` for virtual environment setup in the bundled version, which is generally safe in its controlled context. Data reading operations (`pd.read_csv`, `pd.read_parquet`) handle user-provided paths, which inherently carries some risk, but basic file existence checks are in place. No `eval` or `exec` on user input, and no hardcoded secrets were found. The tool annotations provide good transparency about read/write operations.
Similar Servers
mcp-neo4j
Provides an MCP server to interact with a Neo4j graph database, allowing execution of Cypher read/write queries and schema introspection for graph exploration and generating insights.
memory-graph
Provides intelligent, persistent graph-based memory capabilities for AI assistants and coding agents in Claude Code environments.
mcp-server-datahub
The DataHub MCP Server enables AI agents to discover, explore lineage, understand usage, and generate SQL queries for data assets within a DataHub ecosystem.
Web-Algebra
Translates natural language instructions into RDF operations for loading, querying, and writing Linked Data, consumable via executable JSON or as MCP tools for AI agents.