Real MCP server for the Agentic API Security Governance Framework, exposing tools for customer profile management, record search, ticket creation, and high-privilege operations (delete records, export customers, update billing) with intentionally unenforced scope/identity to demonstrate WAF enforcement patterns.
Server has 6 tools with basic descriptions and parameter schemas, but significant quality gaps. All tools have descriptions (10-60 chars, mostly under the 50-200 char ideal), and input schemas are present with type declarations. However, descriptions lack actionable context for LLM selection, parameter descriptions are minimal or absent, output schemas are undocumented, and error handling provides no recovery guidance. No tool annotations (readOnlyHint/destructiveHint) despite clear risk levels. Three tools marked DESTRUCTIVE/WRITE lack permission gates or confirmation patterns. Tool composition is reasonable (single responsibility), but descriptions do not explain WHEN to use each tool or what happens on failure.
Create a support ticket.
Permanently delete a record. High-privilege — admin scope only.
Export the full customer database. High-privilege — admin scope only.
Fetch a customer's profile by ID.
Search order records by keyword.
Change a customer's billing plan. High-privilege — admin scope only.
Destructive/high-privilege tools (delete_records, export_all_customers, update_billing) lack permission gates, confirmation patterns, or dry-run support. No validation that caller has admin scope before executing.
Tool descriptions are too brief (30-40 chars) and lack actionable context. They do not explain WHEN to use each tool, what it returns, or how it differs from similar tools. LLMs cannot reliably select the right tool.
Output schemas are not documented. Tools return strings (e.g., 'No records matched', 'Created tkt_0001') but LLMs cannot parse structured fields or plan downstream calls. No pagination, no field documentation.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 49 | 2026-07-28+ | v2 |
Parameter descriptions are missing or minimal. 'query' in search_records has a description, but 'customer_id' in get_customer_profile and update_billing lack guidance on format (opaque ID vs email vs name). LLMs must guess.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite clear risk levels. Agents cannot distinguish safe reads from destructive writes without parsing descriptions.