debug
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions in
SKILL.mddescribe workflows that involve processing untrusted external data from media and metadata files. - Ingestion points: The agent is instructed to read content from
edl.json,cut.srt,detect.json, and various media files as specified inSKILL.md. - Boundary markers: The instructions do not provide delimiters or guidance for the agent to distinguish between its own instructions and the content of the external files being processed.
- Capability inventory: The skill utilizes the
Bashtool to executevcut,trx, andffprobe(as referenced inSKILL.md), which are capable of reading and processing the aforementioned files. - Sanitization: No sanitization or validation steps for the input files are mentioned in the debugging procedures.
Audit Metadata