video-editing
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill identifies a workflow where the agent processes external data (transcripts) to generate executable logic.
- Ingestion points: The skill ingests transcripts from long recordings in Layer 2 (Organization) to plan structures and identify segments.
- Boundary markers: No specific delimiters or safety instructions are provided to the agent to distinguish between the transcript data and system instructions.
- Capability inventory: The skill uses the agent to generate and execute FFmpeg shell commands for cutting and concatenating video, as well as Remotion code for rendering visual compositions.
- Sanitization: There is no evidence of sanitization or validation of transcript content before it is used to influence code generation.
- [COMMAND_EXECUTION]: The workflow involves the programmatic generation and execution of shell commands.
- Evidence: The skill provides templates for FFmpeg commands to extract segments, batch cut files, and concatenate video. It also utilizes 'npx remotion' for rendering compositions. This capability, while necessary for the skill's purpose, increases the risk if the upstream planning layer (Claude/Codex) is compromised via indirect injection.
Audit Metadata