MCP server for Google Cloud Storage interactions
Cloud Storage MCP has 6 tools with complete input schemas and descriptions, but quality is uneven. All tools have descriptions (10 - 100 chars, below the 50 - 200 char baseline for LLM optimization). Schemas are present but lack output documentation. Parameter descriptions are minimal (e.g., 'Name of the bucket' is generic). No error handling guidance, no pagination for listFiles/listBuckets despite returning potentially large result sets. No tool annotations (readOnlyHint/destructiveHint). Naming is clear (verb_noun pattern: list, get, upload, download) but parameter naming could be more explicit (e.g., 'file' vs 'file_path'). Security: credentials injected via environment + key files (good), but no audit logging or permission gates visible.
Download a file from a Cloud Storage bucket
Get details of a specific Cloud Storage bucket
Get details of a specific file in a Cloud Storage bucket
List all Cloud Storage buckets in a project
List files in a Cloud Storage bucket
Upload a file to a Cloud Storage bucket
No output schema documentation. Tools return results but LLMs cannot infer response structure (fields, types, pagination). Breaks downstream tool chaining and forces agents to guess field names.
listBuckets and listFiles lack pagination parameters (limit, offset, cursor) and return-count guidance. Large result sets will blow context windows. Baseline: 20 - 50 items max with pagination.
Parameter descriptions are generic and under 50 chars (e.g., 'Name of the bucket'). LLMs need format hints, constraints, and examples. Descriptions should be 50 - 150 chars with actionable detail.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 54 | 2026-07-28+ | v2 |
No error handling guidance. Tools do not document what errors are retryable, user-fixable, or fatal. No recovery hints (e.g., 'Bucket not found, try listBuckets() to see available buckets').
uploadFile accepts 'content' as base64 string but does not document encoding requirement or max size. LLMs may pass raw binary or exceed limits. Add explicit constraint: 'Base64-encoded content, max 5GB'.