caption-word-reveal
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides documentation and configuration parameters for a local video captioning feature. No malicious patterns, obfuscation, or unauthorized access attempts were detected.
- [COMMAND_EXECUTION]: The skill includes instructions for running a local Python script (finish_reel.py) and an ffmpeg command for verification. These operations are standard for media processing workflows and are performed within the local project environment.
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits a potential attack surface for indirect prompt injection through data ingestion. 1. Ingestion points: narration.align.json (processed by finish_reel.py). 2. Boundary markers: Absent. 3. Capability inventory: Local python script execution and ffmpeg video overlays. 4. Sanitization: Not explicitly documented in the instructions. The risk is assessed as low given the local context and the structured nature of the alignment data.
Audit Metadata