groq-common-errors
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the
Bashtool withcurlto interact withapi.groq.com. These operations are used to verify API key validity, list available models, and test payloads, which are legitimate diagnostic actions for the skill's stated purpose. - [INDIRECT_PROMPT_INJECTION]: The skill processes data from Groq API error responses which could serve as a surface for indirect injection if the remote service were compromised. \n
- Ingestion points: Raw JSON error bodies containing
message,type, andcodefields (SKILL.md). \n - Boundary markers: None provided for the error message strings. \n
- Capability inventory: The skill is permitted to use
Read,Grep, andBash(curl:*). \n - Sanitization: There is no explicit sanitization of the error content before the agent interprets it.
- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill requires the
GROQ_API_KEYto be provided via the environment. The instructions demonstrate how to use this key only for authorized requests to the official Groq API service, adhering to secure API integration practices.
Audit Metadata