Loading ToolBench
MCP server exposing RAG-Forge pipeline operations as agent-callable tools
Critical gaps across all dimensions. Tool definitions lack input schemas entirely, no JSON Schema visible for any of the 5 tools. Descriptions are present but minimal (10-50 chars), below the 194-char baseline for production tools. No parameter documentation, no output schema specification, no error handling guidance. Tools are inferred from registration calls rather than explicit schema definitions. This server would not pass code review for production use.
Run evaluation suite and return results
Index new documents into the pipeline
Debug a specific chunk by ID
Execute a RAG query against the indexed pipeline
Return pipeline health metrics and cache stats
No input schemas visible for any tool. LLMs cannot infer parameter types, constraints, or required fields.
Descriptions are 10-50 characters, well below the 194-char baseline. 'Execute a RAG query against the indexed pipeline' lacks WHEN to use it, WHAT it returns, and prerequisites. Descriptions must be 50-200 chars for LLM optimization.
No parameter documentation. rag_status shows Input: {} but other tools' parameters are not visible in source. LLMs cannot determine what parameters to pass or what they control.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 36 | 2026-07-28+ | v2 |
No output schema documentation. LLMs cannot plan downstream tool calls or extract required fields (e.g., chunk_id for rag_inspect, audit_score for rag_audit).
No error handling guidance. If rag_ingest fails (invalid document format, quota exceeded), LLMs have no recovery path. Errors must categorize as retryable, user-fixable, or fatal.