voice-ai-development

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill embeds API keys as string literals and shows them being placed directly into request headers and client constructors, which would require the LLM to handle or output secret values verbatim.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.65). SKILL.md describes a Vapi voice agent setup where runtime customer call events include a live “transcript” from the caller (outsider-authored speech) that can be ingested as LLM input via the agent’s conversation flow/webhooks.