tts

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

• Third-party content exposure detected (high risk: 0.70). The skill explicitly downloads and ingests arbitrary HTTP(S) reference-audio URLs (see prepare_ref_audio in scripts/tts.py, _resolve_reference_audio in scripts/render_timeline.py, and SKILL.md examples like "--ref-audio https://example.com/my_voice.wav"), and those external audio/voice-map inputs are interpreted as part of the TTS/render workflow (controlling per-segment voice, reference audio, and synthesis parameters), so untrusted third-party content can materially influence downstream synthesis behavior.