film-generate-review
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands and local scripts for media processing.
- Evidence: The workflow involves running ffmpeg for video splicing and frame extraction, ffprobe for quality assurance, and local bash scripts (batch.sh, poll.sh) to interface with generative APIs.
- [INDIRECT_PROMPT_INJECTION]: The skill acts as a data processing stage that ingests external URLs.
- Ingestion points: External image URLs are provided as reference parameters in the batch generation command.
- Boundary markers: The skill uses standard command-line flags to delimit these inputs but lacks specific structural boundary markers like XML tags.
- Capability inventory: The agent has the ability to execute shell commands and perform file operations.
- Sanitization: The instructions include a pre-flight checklist for manual verification of asset URLs and Prop/Character references before submission.
Audit Metadata