hyperframes

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

• Third-party content exposure detected (high risk: 0.80). The skill's required caption/transcript workflow (references/captions.md and references/transcript-guide.md) explicitly instructs fetching/transcribing audio via external public APIs (OpenAI/Groq and importing SRT/VTT/YouTube transcripts) and mandates "read the full transcript before choosing style", so the agent ingests untrusted, user- or third-party-generated transcript data that directly drives styling and downstream actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill's sub-composition examples load and execute remote JS at runtime (e.g., the GSAP library at https://cdn.jsdelivr.net/npm/gsap@3.14.2/dist/gsap.min.js), which the skill relies on for timelines/animation, so remote code is fetched and executed during runtime.