ecommerce-visual-copywriting

Pass

Audited by Gen Agent Trust Hub on Jul 10, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill is designed to process user-supplied product data and drafts, which creates an indirect prompt injection surface. This risk is effectively mitigated by a mandatory 'Step 2' pause that requires the user to explicitly confirm extracted selling reasons before the agent proceeds to final content generation. Ingestion points: Step 1 information gathering. Boundary markers: Explicit instructions to pause and wait for user confirmation. Sanitization: Validation of all output against the provided compliance-rules.md library.
  • [SAFE]: The repository includes a dedicated verify-skill.py tool that automatically scans the skill's files for hardcoded secrets, private file paths, and sensitive tokens, demonstrating high security awareness by the developer.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 10, 2026, 06:43 AM
Security Audit — agent-trust-hub — ecommerce-visual-copywriting