media-to-transcript

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

• Third-party content exposure detected (medium risk: 0.30). At runtime, the skill’s pipeline ingests outsider-authored free text from the ASR response by converting the user-supplied media into raw-transcript.md and then building correction-prompt.md from segments[*].text extracted from the downloaded auc-result*.json (result.utterances[*].text / result.text).