oc-cut-judge
Pass
Audited by Gen Agent Trust Hub on Jul 6, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill's primary function is to process and reconcile text-based cut proposals for media editing. All operations are confined to the local project environment.
- [DATA_EXPOSURE]: The skill uses the 'Write' tool to save JSON files to the
<PROJECT>/edl/directory. This is consistent with its stated purpose of generating edit decision lists (EDLs). - [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external 'cut proposers'.
- Ingestion points: Processes 2-3 external proposals for transcript sections.
- Boundary markers: None explicitly defined in the instructions.
- Capability inventory: Uses Bash, Read, and Write tools to generate JSON output.
- Sanitization: No explicit sanitization or validation of input proposals is described, but the output format is strictly defined as JSON, which limits the impact of potential instructions embedded in the input data.
Audit Metadata