extracting-structure
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill defines legitimate video analysis functions that leverage a dedicated 'watch-skill' utility. No attempts to bypass safety filters, escalate privileges, or exfiltrate credentials were identified.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data in the form of video transcripts and OCR text, which represents a potential attack surface for indirect prompt injection.
- Ingestion points: Video content, transcripts, and OCR text processed via subcommands in SKILL.md.
- Boundary markers: The skill does not define specific delimiters or instructions to ignore embedded commands within the video data.
- Capability inventory: The skill uses Bash to execute analysis commands and has Read access to the filesystem.
- Sanitization: No explicit sanitization or input validation for the processed video data is described.
Audit Metadata