openrouter-model-routing
Pass
Audited by Gen Agent Trust Hub on Aug 5, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes user-supplied text to perform task classification and model routing, creating a surface for indirect prompt injection.
- Ingestion points: User input is ingested via the
promptandmessagesparameters in routing functions acrossSKILL.mdandreferences/examples.md. - Boundary markers: The code snippets interpolate user content directly into messages without specialized delimiters or instructions to ignore embedded commands.
- Capability inventory: The skill uses the
openaiSDK to perform network requests for model completions. - Sanitization: No sanitization or filtering is performed on the user prompts before they are forwarded to the model provider.
Audit Metadata