video-editing
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill recommends a workflow where transcript data is used by an agent to generate structured edit decision lists and automation code.
- Ingestion points: Video transcripts provided by the user in Layer 2 (SKILL.md).
- Boundary markers: The current instructions lack explicit delimiters or safety headers to prevent the agent from following instructions hidden within the video transcripts.
- Capability inventory: The skill has the ability to execute shell commands (via FFmpeg), run Node.js rendering tasks (via Remotion), and perform network operations (via ElevenLabs API).
- Sanitization: No sanitization or validation logic is provided to verify the integrity of the generated shell commands or API payloads.
- [COMMAND_EXECUTION]: The skill relies heavily on the execution of FFmpeg shell commands for video manipulation.
- Evidence: Multiple bash snippets for trimming, concatenating, and reframing clips are provided as templates for the agent to follow (SKILL.md).
- Purpose: These commands are necessary for the deterministic cut phase of the described video editing pipeline.
Audit Metadata