ai-video-continuity-director
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides shell commands for creating project directories (
mkdir) and processing video/audio files (ffmpeg). These commands are intended to be run locally by the user or the agent to extract reference frames and audio for maintaining continuity across video shots, which is consistent with the skill's primary function. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external content such as user-provided stories or web novel text to generate video prompts. While this introduces a potential attack surface for indirect prompt injection, the skill employs structured tags (e.g.,
<d>for dialogue) that help delimit user-provided content from the generation instructions. Evidence Chain: 1. Ingestion points: User story/novel text input (SKILL.md Phase 0). 2. Boundary markers: Uses<d>tags for dialogue wrapping. 3. Capability inventory: Local file system access and shell command execution (mkdir,ffmpeg). 4. Sanitization: None explicitly defined.
Audit Metadata