A declarative, secure bridge that exposes internal HTTP APIs to AI agents as typed, allowlisted MCP tools.
Single tool 't_get' has critical definition gaps. Description is trivial (9 chars: 'Get a thing.'), far below the 10-char minimum and 194-char baseline. Parameter 'id' lacks description (null value visible in schema). Schema is present but incomplete, no type info for the parameter itself, only integer type at property level. Tool name is generic and non-specific (does not follow verb_noun pattern clearly, 'get' is good, but 'thing' is vague). No output schema documented. No error handling guidance. The tool appears to be a test fixture (source file is McpServerTests.cs), not a production tool, which explains the minimal effort.
Get a thing.
Tool description is 9 characters ('Get a thing.'), below 10-char minimum and far below 194-char baseline. LLM cannot determine when to use this tool.
Parameter 'id' has null description. LLM cannot infer what 'id' refers to (user ID? resource ID? numeric index?).
No output schema documented. LLM cannot plan downstream tool calls or extract return values.
Tool name 't_get' is non-standard (prefix 't_' is unusual). 'get_thing' is vague, what thing? Should be 'get_<resource_type>'.
No error handling guidance. If the ID is not found, what should the LLM do? Retry? Search? No recovery path documented.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 30 | 2026-07-28+ | v2 |