paw-cra-agent-video-producer
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill relies on system-level CLI tools to perform its core functions, specifically invoking
ffmpegfor video assembly and encoding,egakifor AI generation, andRemotionfor rendering graphics. - [EXTERNAL_DOWNLOADS]: Automatically fetches AI-generated video clips and stock B-roll from well-known services including
fal.aiandPexelsas part of the production pipeline. - [DYNAMIC_EXECUTION]: Employs the
Remotionframework to programmatically generate React components at runtime, which are then rendered into video files via a headless browser or Node.js environment. - [INDIRECT_PROMPT_INJECTION]:
- Ingestion points: Reads external creative briefs, scripts, and subtitle files (SRT/ASS) from the local workspace which may originate from untrusted sources.
- Boundary markers: The instructions do not specify the use of delimiters or specific warnings to ignore embedded instructions within processed scripts.
- Capability inventory: The agent possesses high-impact capabilities including network access via
curl, local filesystem write access, and execution of complex shell commands. - Sanitization: There is no evidence of sanitization or escaping of user-provided script content before it is interpolated into prompts for video and voiceover generation.
Audit Metadata