A CLI tool for scaffolding and analyzing MCP (Model Context Protocol) servers
mcp-farmer is a scaffolding/analysis CLI tool, not a production MCP server itself. The only visible tool definition is 'hello' in src/new/templates/src/server.ts, which is a template example, not a real tool. The 'hello' tool has minimal description (19 chars: 'Says hello to the provided name'), a single parameter 'name' with basic schema, and no documented output schema. The tool lacks detail on return format, error handling, and proper context for LLM selection. As a template generator, mcp-farmer provides scaffolding but the example tool definition falls well short of production quality. The codebase includes vetting logic (src/vet/tools.ts) that checks for description quality (minimum 50 chars), missing input descriptions, dangerous/PII keywords, and similar tools, but this is analysis code, not evidence of the scaffolded servers meeting those standards.
Says hello to the provided name
Tool description is critically short (19 chars), fails baseline of 50+ chars per mcp-farmer's own vetting rules
No documented output/return schema, LLMs cannot infer what 'hello' returns or chain downstream calls
Tool name 'hello' does not follow verb_noun pattern, should be 'greet_user' or similar to signal action to LLM
No error handling documentation, what happens if 'name' is empty, null, or excessively long? How should LLM recover?
Input parameter 'name' description is generic ('Name to greet'), lacks format constraints, length limits, or examples of expected input
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 38 | 2025-06-18+ | v2 |
| 2026-03-09 | F | 21 | - | v1 |