speech-summarization
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted media files (audio/video), which creates an attack surface where malicious spoken content could attempt to influence the AI's summary or downstream chained actions. \n
- Ingestion points: Media files are read from the filesystem via
readFileinexamples/node.md. \n - Boundary markers: No explicit delimiters or warnings to ignore embedded instructions are used in the provided implementation. \n
- Capability inventory: The skill reads local files and performs network POST requests to Zoom's API. \n
- Sanitization: No sanitization or filtering is applied to the media input or the generated summary output. \n- [SAFE]: The skill uses official Zoom API endpoints (
api.zoom.us) and adheres to secure development practices, such as discouraging client-side exposure of API keys and using standard environment variable configuration.
Audit Metadata