MCP server for Power Platform and Dataverse developer operations, integrated with VS Code extension. Provides tools for authentication, connection reference analysis, solution management, metadata browsing, and plugin registration for AI agents.
This MCP server has severe definition quality gaps. Only 2 tools are exposed, both with minimal descriptions (under 100 chars) and no visible input schemas in the provided source code. The tool names follow verb_noun convention (ppds_auth_who, ppds_connection_references_analyze), but without access to actual parameter definitions, input validation rules, or detailed output schemas, the implementation appears incomplete. The descriptions provided are functional but lack the depth needed to guide LLM tool selection (e.g., 'Get the current authentication profile context...' doesn't explain when to call this vs alternatives, or what happens if no context exists). No evidence of enum constraints, parameter validation, error recovery guidance, or output schema documentation. The server appears to be in early development.
Get the current authentication profile context including identity, connected environment, and token status. Use this to understand which Dataverse environment queries will run against.
Analyze connection references for orphaned references (not used by any flow) and orphaned flows (referencing missing connection references). Useful for deployment cleanup.
No input schemas visible in source code. Both tools have schema score of 0 because parameter definitions, types, and constraints are not present in the provided codebase excerpt.
Tool descriptions are too brief (<100 chars) and lack context on when to use each tool, prerequisites, and error cases. 'Get the current authentication profile context...' doesn't explain failure modes or how this relates to the other tool.
No documented output schemas. LLMs cannot plan downstream calls or extract relevant fields if the response structure is unknown. What fields does ppds_auth_who return? Is there a token_status field? An environment_id?
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 39 | 2026-07-28+ | v2 |
| 2026-03-09 | D | 54 | - | v1 |
No visible parameter descriptions or type information in the source provided. Cannot verify if parameters accept enums, have defaults, or constraints. This violates the requirement that every parameter have a description and type.
No error handling guidance. If ppds_auth_who fails (e.g., no active profile), what should the LLM do? Should it suggest creating a profile, or is this a fatal condition? No recovery paths documented.
Tool naming could be more discoverable. 'ppds_connection_references_analyze' is 33 characters and uses an underscore-prefixed namespace. The verb 'analyze' is less specific than 'find_orphaned_connections' or 'detect_unused_references'. Current name requires reading the description to understand intent.
No pagination or limit handling documented. If ppds_connection_references_analyze returns hundreds of orphaned items, is there a way to limit results or page through them? Large unbounded responses will blow context windows.