OpenAI-compatible multi-account AI gateway with MCP server support, credential management, quota tracking, and control-plane integration
Critical failure: Tool definitions are inferred from file references only. No explicit tool registration, schemas, or parameter definitions visible in provided source code. Tool names and descriptions appear only in comments/test files, not in actual MCP server implementation. Cannot verify input schemas, output schemas, or parameter types.
Consume a reset credit for an upstream provider account
Discover available models from an upstream provider account
List available reset credits for an upstream provider account
Read usage statistics from an upstream provider account
Validate credentials for an upstream provider account
No input schemas visible in source code. Cannot verify parameter types, constraints, or descriptions. Hard cap rule applies: schema score must be 0 for all tools.
Tool definitions are inferred from file paths only (server_test.go, adapter.go). No explicit tool registration visible in MCP server code. Cannot verify actual implementation.
Descriptions are single-sentence summaries (20-50 chars). Rubric requires 10-1024 chars with WHAT/WHEN/WHY context. Current descriptions lack parameter guidance, prerequisites, and return value documentation.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 29 | 2026-07-28+ | v2 |
No output schema documentation visible. LLMs cannot plan downstream calls or extract required fields (e.g., model IDs, credit amounts, usage metrics).
No error handling guidance visible. Tools lack recovery hints, retryability classification, or actionable error messages per pattern:recovery-guide.