transcribe-refiner

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). This skill enforces "zero content loss" and explicitly preserves/code-snippets/commands and a Topic Inventory, so any secrets (API keys, tokens, passwords) spoken or present in the transcript would be retained and output verbatim, creating a direct exfiltration risk.

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly ingests and processes raw caption files and meeting transcripts from open/public third-party sources (e.g., "Zoom, YouTube, Teams, Google Meet, Otter.ai" and user-provided .txt/.vtt/.srt or raw paste) and is designed to read, interpret, correct, and preserve their substantive content (including names, URLs, and code) as part of its workflow, creating a clear vector for indirect prompt injection.