groq-sdk-patterns
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides documentation and code snippets for using official SDKs (groq-sdk, groq) from a well-known service provider. No malicious code, exfiltration, or obfuscation was detected.
- [INDIRECT_PROMPT_INJECTION]: The provided code patterns for chat completions establish an ingestion surface for untrusted data that will be processed by an LLM. * Ingestion points: The messages parameter in the complete and streamCompletion functions within references/typescript-patterns.md and references/python-patterns.md allows external data to enter the model context. * Boundary markers: None are present in the simplified code examples to separate system instructions from user content. * Capability inventory: The patterns facilitate network requests to the Groq API (api.groq.com) using the provided SDKs. * Sanitization: The examples do not include explicit input sanitization or validation of the message content before processing.
Audit Metadata