A Python-based MCP server for generating SEO metadata (alt text) from product images
This server has a single tool 'generate_alt' with severe definition quality issues. The tool name lacks a clear action verb (should be 'generate_alt_text' for clarity), the description is minimally adequate (59 chars, below the 194 char production baseline), parameter descriptions are completely absent, and the output schema is not documented. The implementation returns a hardcoded string rather than actual image processing, indicating this is a stub/mock. No error handling, no input validation, and no security considerations are evident. The server would not pass basic code review.
Generate SEO-optimized alt text for product images
Tool name 'generate_alt' is ambiguous, should be 'generate_alt_text' to clearly indicate it generates alternative text. Current name omits the object being generated.
Parameter 'image_url' lacks a description entirely. LLMs cannot infer whether this is a public URL, requires authentication, supports specific formats, etc.
Output schema is not documented. The tool returns a dict with 'alt_text' field, but LLMs have no visibility into this structure. Production tools must document return types explicitly so agents can plan downstream use.
Tool description (59 chars: 'Generate SEO-optimized alt text for product images') lacks context on WHEN to use it vs alternatives, what formats are accepted, or what the output looks like. Production baselines average 194 chars.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 24 | 2026-07-28+ | v2 |
No input validation or error handling. The tool returns a hardcoded response regardless of input. No guidance on what happens if image_url is invalid, unreachable, or unsupported format.
Implementation is a hardcoded stub returning the same text for any input: return {"alt_text": "High quality modern athletic shoe product photo."}. This is not a functional tool, it does not actually process images or generate context-aware alt text.
No indication of image format support (JPEG, PNG, WebP, SVG?), size limits, or required image properties. Parameter description must state constraints.
No mention of whether the tool is idempotent. Repeated calls with the same image_url should produce identical alt text (they do, due to hardcoding, but this is accidental not by design).