Local model router / proxy across Ollama, vLLM, and cloud with fallback
Single tool 'modelroute_scan' has a basic description but critically lacks parameter descriptions and output schema documentation. The input schema defines only 'target' as a string with minimal context. No error handling guidance, no output structure documented, and no indication of what 'JSON findings' contains. The description is vague about the tool's actual purpose (is it a code scanner? a model router? both?). Naming is acceptable but descriptions fall far short of production standards.
Local model router / proxy across Ollama, vLLM, and cloud with fallback. Returns JSON findings.
Parameter 'target' lacks description. LLM cannot infer whether it expects a file path, directory, URL, or model name.
Output schema not documented. Tool returns 'JSON findings' but structure is unknown, no field names, types, or examples provided.
Tool description is ambiguous. Mentions 'model router / proxy' and 'fallback' but also 'scan', unclear whether this scans code, routes LLM requests, or both.
No error handling guidance. What happens if target path doesn't exist? Is it retryable? What should the LLM do?
No constraints on 'target' parameter. Should it be a relative path, absolute path, or URL? What's the max length? Are symlinks followed?
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 42 | 2026-07-28+ | v2 |