the-checkout-auditor

Pass

Audited by Gen Agent Trust Hub on Aug 14, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious code or risky behaviors were detected. The skill follows security best practices by defining clear boundaries between instructions and untrusted data.
  • [PROMPT_INJECTION]: The skill processes untrusted external content (checkout pages, screenshots), representing a potential indirect prompt injection surface. Ingestion points: Checkout walkthrough, policy details, and funnel metrics identified in SKILL.md. Boundary markers: Explicit instructions to treat external text as data and never instructions are present in SKILL.md and references/agent-security.md. Capability inventory: The skill does not possess dangerous tools like arbitrary command execution or file writes. Sanitization: Mandatory HTML-escaping and URL validation rules are defined in the security references.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 14, 2026, 12:03 PM