cutagent-opening-hooks
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes source material to generate video hooks and references local execution via the CutAgent CLI. This creates a surface where malicious instructions embedded in the source material could potentially influence the agent's behavior during execution.\n
- Ingestion points: Source material (video transcripts, scripts, or content) used as the basis for the hook design as described in the "Find the hook" section.\n
- Boundary markers: None identified; the instructions do not provide specific delimiters or "ignore-instruction" warnings for the source data.\n
- Capability inventory: Mentions local execution via CutAgent CLI, SDK interaction, and the creation of editable Fusion graphs and .setting files.\n
- Sanitization: No sanitization or validation of the source content is specified.
Audit Metadata