SupraWall security gateway as an MCP server
Three tools with basic schemas and descriptions, but significant gaps in parameter documentation, output schema clarity, and error handling guidance. check_policy and request_approval have incomplete parameter descriptions (context and reason lack detail on format/constraints). log_action is partially visible in source (truncated in index.ts). No tool describes what it returns or how to use results downstream. Descriptions are adequate length (50-100 chars) but lack actionable detail on when/why to call each tool. No error recovery guidance. Security concern: API keys passed as parameters in server.ts (line with apiKey in request body).
Check if an AI action complies with configured compliance policies
Log an action to the audit trail
Request human approval for a potentially sensitive action
API key exposed in tool parameters. server.ts passes apiKey in request body to external API. Credentials must use server-side injection only.
Output schemas not documented. Tools return JSON responses but LLMs cannot plan downstream calls without knowing field names/types. check_policy returns decision/reason/risk_score but this is not in the tool definition.
Parameter 'context' in check_policy lacks description of expected structure/format. LLMs cannot infer what fields to include.
No error recovery guidance. Tools catch errors and return isError=true with raw API responses, but do not tell LLM what to do next or whether to retry.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 47 | 2026-07-28+ | v2 |
log_action source code truncated in index.ts. Cannot verify complete schema or implementation. Tool definition in server.ts is missing entirely.