MCP server that provides tools for finding overlaps between software tools and discovering cost savings through tool consolidation
Single tool 'find_overlaps' has a reasonable name and description (89 chars), but critical gaps undermine quality. Input schema is visible with array type and description, but lacks critical constraints: no enum for tool names, no format validation, no min/max array bounds. Parameter description mentions examples ('HubSpot', 'Salesforce') which violates the pattern of using enums instead. Output schema is completely undocumented, the client code shows result.content[0]?.text but no formal schema is declared. No error handling guidance, no pagination support despite potentially returning large overlap matrices, no per-tool success/failure breakdown. The tool does one thing (find overlaps) which is good, but the definition lacks the rigor expected of production tools.
Find overlaps between software tools and discover cost savings from tool consolidation
Output schema completely undocumented. Client code shows result.content[0]?.text but no formal schema definition exists. LLMs cannot plan downstream operations or extract structured data.
Input parameter 'tools' lacks constraints. No enum of valid tool names, no min/max array length, no format validation. LLMs will hallucinate invalid tool names.
Parameter description includes example values ('HubSpot', 'Salesforce', 'Outreach', 'ZoomInfo') which LLMs tend to reuse literally. Should use enum constraint instead.
No error handling guidance. If a tool name is invalid or overlap data unavailable, LLM receives no recovery path. Should include 'Try search_tools() first' or similar.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 62 | 2026-07-28+ | v2 |
No pagination or result limits documented. If overlap matrix is large, response could exhaust context window. Should cap results and offer pagination.