assemblyai-sdk-patterns

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md's Step 2 Transcription Service explicitly calls client.transcripts.transcribe({ audio }) and returns the transcribed text, which means the skill ingests arbitrary external audio URLs (third-party/untrusted content) and exposes resultant text that the agent is expected to read and could influence subsequent actions.