Exposes KMS operations as Model Context Protocol tools so Claude and other MCP-speaking LLMs can act as operators within the scope of a Principal.Agent
This server is in early scaffolding stage. The McpServer.scala file contains only a placeholder list of tool names as strings with no actual tool registration, schemas, parameter definitions, or descriptions visible in the source. The tools are declared as a List[String] with no MCP protocol integration. No input schemas, output schemas, or parameter documentation exist. The brief one-line descriptions in the tool list ('List cryptographic keys', 'Create a new cryptographic key', etc.) are present but lack the depth required by the rubric (10-1024 chars, with parameter-level documentation). No error handling, validation, or security patterns are evident. This is a pre-alpha implementation.
Search audit logs
Create a new cryptographic key
Destroy a cryptographic key (step-up required)
List cryptographic keys
Revoke a cryptographic key (step-up required)
Rotate an existing cryptographic key
Get policy recommendations
No input schemas visible. All 7 tools lack JSON Schema definitions for parameters.
No parameter descriptions. Tools lack per-parameter documentation explaining what each input controls, expected format, constraints, or valid values. LLMs cannot infer parameter meaning from names alone.
No output schemas documented. No specification of what fields each tool returns, their types, or structure. Agents cannot plan downstream calls or extract required data.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 37 | 2026-07-28+ | v2 |
Destructive operations (aegis.keys.destroy, aegis.keys.revoke) lack confirmation/dry-run patterns. No evidence of step-up verification or recovery guidance despite 'step-up required' risk labels.
No error handling or recovery guidance visible. No indication of how tools communicate failures, retryability, or next steps to the agent.