Exploring MCP intent clusters

Intent clustering takes the free-text $mcp_intent values agents attach to their tool calls, embeds them, and groups semantically similar goals into clusters. Attribution is per call: each call is credited to its own intent (calls without one inherit the most recent prior intent in the same session), so a tool's counts reflect the intent it actually served. Each cluster carries its tool distribution, call counts, and error rates — answering "what are people trying to do, and does it work?" rather than "which tool was called". The snapshot also carries a tool-centric pivot answering the reverse question: for a given tool, which intents drive its usage, how often do agents find it, and who does it compete with.

Unlike tool quality and sessions (which ultimately aggregate $mcp_tool_call), clustering needs embeddings and is not expressible in SQL. It is served by two typed tools backed by a stored snapshot.

Tools