MCP server for checking real-time API and cloud service status via API Status Check
This server has solid naming conventions (all tools start with action verbs: check_, list_) and complete input schemas with Zod validation. However, tool descriptions are sparse and under 100 characters on average, lacking context about when to use each tool, what they return, and their relationships. Parameters lack descriptions in their Zod definitions, while the server uses .describe() for some (e.g., 'API slug, e.g. cloudflare'), others like check_url's format='uri' constraint is expressed only via Zod without explanation text. Output schemas are not documented, responses are coerced to JSON text via asText() and wrapped in a generic {content: [{type: 'text', text: ...}]} envelope, which forces LLMs to parse unstructured JSON strings rather than work with typed fields. Error handling returns a generic 'Error: {message}' with isError flag but no guidance on recovery steps or categorization (retryable vs fatal). The server handles 5 read-only tools for a public API status check service, so destructiveness is not a concern, but lack of output documentation and error guidance limits agent effectiveness.
Check the status of all APIs in a specific category
Check the status of a specific API by slug
Check the status of a public URL
List all available APIs with their current status
List all available categories of APIs
Output schemas not documented. Tools return JSON wrapped in {content: [{type: 'text', text: '...'}]}, forcing LLMs to parse unstructured text rather than typed fields. Response fields (status, apis, categories, errors) are not described.
Tool descriptions too brief (10-50 chars). E.g., 'List all available APIs with their current status' is good, but missing: what fields are returned? Why choose this over check_status? What limits apply? Descriptions should be 50-200 chars for LLM planning.
Error handling is generic. Errors return 'Error: {message}' with isError flag, but no categorization (retryable/user-fixable/fatal) or recovery guidance. E.g., if API is down, the LLM has no hint to retry later or fallback to cached data.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | C | 63 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 45 | - | v1 |
No pagination or result limits. list_apis returns all APIs without limit or offset. If the status check service returns 1000+ APIs, the full response bloats the context window and wastes tokens.
check_url parameter uses Zod .url() constraint but no human-readable description of the constraint or acceptable formats. LLMs may not understand the URI format requirement without explicit guidance.