gen-ai-persona-creation

Pass

Audited by Gen Agent Trust Hub on Jun 25, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill invokes the 'gen-ai' command-line interface to interact with various generative models. It uses shell pipelines with 'grep' and 'head' to process JSON output and extract asset URLs for local storage.
  • [EXTERNAL_DOWNLOADS]: The skill uses 'curl' to download generated images and videos from remote URLs. These downloads are associated with the intended functionality of retrieving assets from the AI generation service.
  • [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by incorporating user-provided briefs into generative prompts.
  • Ingestion points: User-provided natural language descriptions used to seed persona creation (Step 1).
  • Boundary markers: No explicit delimiters or boundary markers are utilized in the prompt templates shown in the procedures.
  • Capability inventory: The agent possesses the capability to execute shell commands ('gen-ai', 'curl') and write files to the local directory structure.
  • Sanitization: No specific validation or escaping mechanisms are described for the natural language input before its interpolation into model prompts.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 25, 2026, 12:52 AM
Security Audit — agent-trust-hub — gen-ai-persona-creation