openrouter-sdk-patterns
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides code patterns that ingest user-controlled strings (prompts and system instructions) to be processed by a Large Language Model. While this is the standard functionality of such a wrapper, it establishes an attack surface for indirect prompt injection if the inputs contain malicious instructions.
- Ingestion points: The 'prompt' and 'system' parameters in the 'OpenRouterClient.complete' (SKILL.md) and 'OpenRouterClient.chat' (references/examples.md) methods.
- Boundary markers: None are implemented in the provided code snippets to distinguish between instructions and data.
- Capability inventory: The code is authorized to perform network requests to the well-known OpenRouter API service via the OpenAI SDK.
- Sanitization: No sanitization or filtering logic is included in the reference implementations.
Audit Metadata