mcp-compressor wrapper proxy in front of a local-model agent
Parent: Mac local LLMs: Agent clients, context and compaction · Published reference · snapshot 2026-10-05
↓ Facts as markdownall context files
The model first sees a small compressed surface, asks for the full schema of one tool, then invokes it.
These notes link each claim to its source. Some citations identify archived research reports that are no longer hosted on this site. A published reference means the content is available; it does not certify independent review or accuracy.Read the editorial policy and follow the sources before relying on a claim.
Facts
- The model first sees a small compressed surface, asks for the full schema of one tool, then invokes it. [source]
- Pattern 1 is a wrapper in front of any MCP client; pattern 2 is a local session proxy for SDKs and generated clients, which call it with a session token and need no stdio subprocess. [source]
- No source found measures a local small model choosing tools through the proxy; the existing open question stands. [source]
- Task success of a 9B to 35B local model through get_tool_schema then invoke_tool round trips. [source]
- The mcp-compressor docs describe a wrapper pattern and a local session proxy pattern with session tokens [source]
- No source read gives a measurement of mcp-compressor with a local model; the claim stays open [source]
Children
- No children recorded.