framevideo-ai-production

Pass

Audited by Gen Agent Trust Hub on Jul 21, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the npx framevideo command to retrieve available AI models and manage tasks on the vendor's platform. These commands are integral to the skill's function and target vendor-owned infrastructure.
  • [EXTERNAL_DOWNLOADS]: The workflow involves downloading generated images and videos from the Chanjing AIGC platform to local directories (assets/ai-creation/). These operations are consistent with the skill's stated purpose of asset production.
  • [PROMPT_INJECTION]: The skill processes untrusted user scripts to generate prompts with the following evidence chain: 1. Ingestion points: Raw scripts provided via SKILL.md. 2. Boundary markers: Explicit negative constraints (e.g., 'no captions', 'no dialogue changes') are appended to outputs. 3. Capability inventory: Execution of npx framevideo and local project file writes. 4. Sanitization: Instructions mandate literal preservation of original dialogue and character names to prevent manipulation.
  • [DATA_EXFILTRATION]: The skill maintains local metadata files in the .framevideo/ directory to track task status. No transmission of sensitive information to unauthorized third-party services was identified.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 21, 2026, 06:25 AM
Security Audit — agent-trust-hub — framevideo-ai-production