MCP server for image processing, metadata extraction, and format conversion with Azure Blob Storage integration
This server presents a mixed quality picture. It has 6 tools across two transports (STDIO and HTTP). Tool names follow the verb_noun pattern (get_, convert_, list_, shrink_). Descriptions are present but vary in quality, some are clear (get_image_metadata), others are repetitive or misleading (list_blob_image_metadata and list_storage_accounts have identical descriptions despite different purposes). Input schemas are visible and properly typed. However, parameter descriptions are sparse or missing detail, error handling guidance is weak, and no output schemas are documented. The STDIO transport tools have good C# attribute-based registration with [McpServerTool] markers, but the HTTP tools show only partial source code, making full validation impossible. Tool composition is reasonable, each tool has a single responsibility, but chaining between tools (e.g., list_storage_accounts → shrink_blob_images) lacks guidance in descriptions. Security is a concern: the Azure tools accept subscriptionId and storageAccountName as parameters without validation guidance, increasing injection risk.
Convert an image to WebP format with configurable quality and return metadata for the new file.
Get image metadata including height, width, and EXIF data if it is available.
Retrieves a list of Azure storage accounts, containers, and blobs in a region or subscription.
List all deleted large blob images by container. Returns container information with a list of deleted blobs per container.
Retrieves a list of Azure storage accounts, containers, and blobs in a region or subscription.
Convert blob images to a smaller format (WebP). The original image can be optionally deleted.
Duplicate/misleading descriptions: list_blob_image_metadata and list_storage_accounts both claim to 'Retrieve a list of Azure storage accounts, containers, and blobs in a region or subscription.' This violates naming clarity, the descriptions do not distinguish when to use each tool. LLMs will conflate them.
Parameter descriptions lack detail and constraints: subscriptionId and storageAccountName in Azure tools have no format, length, or validation guidance. Required parameters (e.g., storageAccountName in list_blob_image_metadata) are not marked as required in descriptions. No guidance on how to obtain these values or when they are optional.
No output schemas documented: None of the 6 tools document their return types or output structure. The STDIO tools return ImageMetadata and ConvertedImageMetadata (visible in source), but Azure tools' return structures are inferred only. LLMs cannot plan downstream calls without knowing what fields to expect.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | D | 58 | <=2025-11-25 | v2 |
| 2026-03-09 | D | 51 | - | v1 |
Error handling lacks recovery guidance: The STDIO tools throw McpException and FileNotFoundException with minimal context. The HTTP tools show incomplete error handling in partial source code. No errors guide LLMs on retry strategy, user fixes, or fallback tools.
Destructive tool (shrink_blob_images) with deleteOriginal flag defaulting to false, but no confirmation or dry-run support. LLMs can trigger permanent data loss without safeguards or warnings in the description.
Tool composition gap: list_storage_accounts and list_blob_image_metadata return different result structures, but no chaining guidance. An LLM using list_storage_accounts to get a storageAccountName for list_blob_image_metadata has no assurance the returned names match the parameter format.