Single tool 'generate-llms' with minimal definition quality. The tool has a basic description and a single string parameter with minimal schema documentation. The tool description is functional but vague about context and when to use it. Parameter lacks guidance on expected format or constraints. No output schema is documented. The tool appears to concatenate a prompt file with user input and return it as text, but this is inferred from code review rather than explicit in the definition. No error recovery guidance. Critical gaps in parameter descriptions, output schema documentation, and error handling patterns.

• Add explicit output schema: document that the tool returns {content: [{type: 'text', text: string}]} and explain what the text contains (a prompt template, a concatenated prompt, etc.).
• Add description to 'prompt' parameter: e.g., 'User requirement text describing the desired llms.txt structure and content (no length limit specified; typically 50-500 characters)'.
• Expand tool description to ~100-150 characters with concrete semantics: 'Returns a system prompt for generating llms.txt documentation files. The returned prompt can be passed to an LLM to generate the actual documentation. Does not write files directly; use alongside a file-writing tool.'
• Document error cases: 'If the prompt template file (generate.md) cannot be loaded, returns an error message. Users should verify the template file exists.'
• Rename tool from 'generate-llms' to 'create_llms_prompt' or 'get_llms_generation_prompt' to match verb_noun convention and clarify that it returns a prompt, not the final files.
• Consider adding parameters for customization: allow the caller to specify output directory, file naming, or content filters rather than hard-coding them.
• Add output pagination note if the prompt is expected to be large: document any truncation or limits.