groq-inference

Pass

Audited by Gen Agent Trust Hub on Mar 14, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted data such as user-defined prompts and external image URLs, which are directly interpolated into Groq API requests. This creates a potential surface for indirect prompt injection where instructions embedded in the data could influence the agent's behavior.
  • Ingestion points: Data enters the agent context through the prompt and image_url parameters used in SKILL.md, templates/groq-client.py, and reference/vision-multimodal.md.
  • Boundary markers: The provided code examples lack explicit delimiters or system instructions to isolate data from model instructions.
  • Capability inventory: The skill facilitates network communication with the Groq API and performs file system read operations for processing audio and image data.
  • Sanitization: No validation or sanitization mechanisms are implemented for the external data passed to the inference models.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 14, 2026, 01:42 AM