speech-to-text

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.90). Most links are documentation and media on a single domain (inference.sh), but the quick-start instructs piping a remote install script (curl https://cli.inference.sh | sh) — a high‑risk pattern because executing an unverified remote shell script can deliver malware.

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly ingests arbitrary public audio/video URLs (e.g., the examples using "audio_url": "https://audio.mp3" and "video_url": "https://video.mp4" in SKILL.md) and transcribes that untrusted, user-provided content as part of its workflow, which could allow instructions embedded in third-party audio to influence downstream actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The Quick Start instructs running "curl -fsSL https://cli.inference.sh | sh", which fetches and executes remote installer code from https://cli.inference.sh at runtime and is a required dependency for running the infsh apps.