Reproducible AI agent and MCP security profiles, historical drift, badges, datasets, and APIs
This is a frontend/CLI tool for security profiling, not an MCP server. The single tool 'select_security_profile' is defined in a React component (app/page.tsx), not via MCP protocol. No MCP server implementation is visible in the source. The tool has a reasonable schema with enum constraints and a clear description, but the codebase is a Next.js/Vite frontend application with Python CLI utilities, not an MCP server. Tool definition appears inferred from UI code rather than explicit MCP registration.
Select one controlled AgentSec profile by its stable ID and update the visible evidence detail panel.
Not an MCP server. Source is a Next.js frontend + Python CLI, not an MCP protocol implementation. No MCP server code (initialize, tools/list, tools/call handlers) found.
Tool definition inferred from React component, not explicitly registered via MCP. Cannot verify actual MCP schema, input validation, or error handling.
Parameter 'profileId' lacks description. Schema shows enum values but no explanation of what each profile represents or when to select each.
No output schema documented. Tool description does not specify what 'update the visible evidence detail panel' returns or what fields the agent receives.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 40 | 2026-07-28+ | v2 |
No error handling guidance. No documentation of what happens if an invalid profileId is passed or if the profile cannot be loaded.