specialist-marketing-cross-border-ecommerce

Pass

Audited by Gen Agent Trust Hub on Aug 15, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill defines a prompt for a marketing sub-agent using the Agent tool. No malicious code, obfuscation, or suspicious patterns were detected in the instructions or metadata.
  • [PROMPT_INJECTION]: The skill accepts user input through the {the user's question} placeholder, which is a standard interface for agent-based skills. The absence of high-risk capabilities like file system access, network operations, or arbitrary command execution minimizes the risk associated with this injection surface.
  • Ingestion points: The user input is interpolated into the Agent tool prompt in SKILL.md via {the user's question}.
  • Boundary markers: None are explicitly defined around the user input placeholder.
  • Capability inventory: The skill is restricted to using the Agent tool to spawn a sub-agent. No subprocess, file system, or network capabilities are present.
  • Sanitization: No input sanitization or filtering logic is specified in the skill definition.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 15, 2026, 08:50 PM
Security Audit — agent-trust-hub — specialist-marketing-cross-border-ecommerce