Collection of MCP servers for HPE GreenLake platform APIs including audit-logs, devices, reporting, service-catalog, subscriptions, users, and workspaces
This server has two tools with identical names ('invoke_dynamic_tool'), which violates basic tool composition and naming clarity principles. Both tools have identical parameter schemas and descriptions that are generic and lack specificity about what API endpoints they target. The descriptions mention 'audit-logs API' and 'devices API' but this distinction is not reflected in the tool names themselves, creating severe ambiguity for LLM selection. Parameter descriptions are present but minimal (under 50 chars for most). No output schema documentation is visible in the provided source code. The tools appear to be generic API invocation wrappers rather than domain-specific, task-focused tools aligned with user intent.
Executes any audit-logs API endpoint dynamically with parameter validation and schema support
Executes any devices API endpoint dynamically with parameter validation and schema support
Duplicate tool names: Both tools are named 'invoke_dynamic_tool', making them indistinguishable to LLMs. The server description indicates one targets 'audit-logs' and one targets 'devices', but this is not reflected in the tool names.
Generic, undifferentiated tool names: 'invoke_dynamic_tool' does not follow verb_noun convention (e.g., 'get_', 'create_', 'list_'). It reads as a generic wrapper rather than a domain-specific action. No indication of what business problem is being solved.
Minimal, generic descriptions: Both tool descriptions are only ~100-120 characters and provide minimal guidance on WHEN to use each tool. No explanation of prerequisites, expected outcomes, or how to construct endpoint_identifier values correctly.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 39 | <=2025-11-25 | v2 |
| 2026-03-09 | F | 14 | - | v1 |
No visible output schema documentation: The source code snippet does not show what these tools return. Without documented output schemas, LLMs cannot plan downstream operations or extract required fields for chaining.
Parameter format ambiguity: The 'endpoint_identifier' parameter description says 'METHOD:PATH format (e.g., GET:/api/v1/pets)' but provides only one example. No mention of how path parameters like {petId} should be provided in relation to the 'parameters' dict. This invites format errors.
Generic wrapper pattern: The tools appear to be generic API invocation bridges rather than task-focused tools aligned with user intent. Users do not think in terms of 'invoke a dynamic endpoint', they think 'list audit logs' or 'get device details'. Generic wrappers force LLMs to reason about API structure and construct requests, rather than focusing on business outcomes.
No per-parameter validation guidance: The 'validate_schema' boolean defaults to true, but there is no description of what happens if validation fails, what error messages are returned, or how to fix invalid parameters. This violates the error-guidance pattern.
Missing parameter descriptions: The 'parameters' parameter description is minimal ('Request parameters (path and query parameters only)') and lacks examples of valid structures, how to specify nested values, or what happens with malformed input.