Model Context Protocol server for generating QR codes
Single-tool server with well-defined schema and clear descriptions. The generate-qrcode tool has a complete input schema with proper types, enums, and descriptions. Tool naming follows verb_noun pattern (generate-qrcode). Descriptions are adequate but could be more LLM-optimized. No output schema documentation visible in provided source. Error handling and recovery guidance not evident. Resource-based architecture (qrcode outputs as resources) is well-suited to the domain but adds complexity without clear composition patterns.
Generate QR codes in various formats with customizable error correction levels and sizes
Output schema not documented. Tool description states what formats are available (image, text) but does not specify the structure of the response object (e.g., does it return a resource URI, base64 data, text representation?).
Tool description lacks recovery guidance for error cases. No indication of what happens on invalid input (e.g., oversized content, invalid correction level) or how the agent should respond to failures.
Parameter descriptions are present but minimal. 'content' description is just one line; no guidance on maximum length, character restrictions, or encoding. 'size' parameter has range but no explanation of what the numeric range means in practical terms (pixels? dimension scale?).
No indication of when to prefer 'image' vs 'text' format or when one is more suitable. LLM must guess based on context rather than explicit guidance in the description.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-21 | C | 63 | 2026-07-28+ | v2 |
| 2026-03-09 | D | 54 | - | v1 |
Tool is marked as READ_ONLY but generates and stores outputs as resources. This may create confusion about side effects, generating a QR code creates a new resource, which is a state change even if it's read-only in terms of not modifying existing data.