agency-cross-border-e-commerce-specialist
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill describes workflows involving the ingestion of external data, such as product reviews and reports from tools like Jungle Scout and Helium 10. While this creates a theoretical attack surface for indirect prompt injection, the skill does not include any executable code, file-writing capabilities, or network operations that could be exploited. This is standard functionality for an e-commerce analysis agent.
- [REMOTE_CODE_EXECUTION]: No remote code execution patterns, script downloads, or external package installations were detected. The skill is composed entirely of instructional markdown.
- [DATA_EXFILTRATION]: No sensitive file access or network exfiltration patterns were found. Mentions of external platforms (e.g., Stripe, PayPal, Amazon) are related to business strategy and do not involve credential harvesting or unauthorized data transmission.
- [OBFUSCATION]: No hidden content, encoded strings (Base64, hex), or homoglyph attacks were detected within the documentation or metadata.
Audit Metadata