vlog-cut-pipeline
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill orchestrates a series of shell commands (
vlog-cut-tts,vlog-cut-index,vlog-cut-plan,vlog-cut-render, etc.) to process video assets, generate narration, and assemble a final rough cut.\n- [EXTERNAL_DOWNLOADS]: The instructions include a check for thewhispercommand and suggest installing the officialopenai-whisperpackage via pip if it is missing.\n- [INDIRECT_PROMPT_INJECTION]:\n - Ingestion points: The skill reads user-provided text from
script.jsonand asset metadata fromassets_index.json(SKILL.md).\n - Boundary markers: None explicitly mentioned for the script content processing to distinguish instructions from data.\n
- Capability inventory: Execution of local CLI tools and file system operations within the specified project directory.\n
- Sanitization: Includes a dedicated validation step (
vlog-cut-validate) to ensure the generated timeline is consistent with the source assets and timing.
Audit Metadata