The Agent Skills Directory

[SAFE]: The skill content is educational and instructional, focusing on legitimate business commerce implementations. A thorough analysis of all 11 threat categories—including prompt injection, data exfiltration, obfuscation, and privilege escalation—showed no evidence of malicious behavior or security policy violations.\n- [INDIRECT_PROMPT_INJECTION]: The skill describes systems that ingest business data (e.g., buyer notes in quote workflows). While these represent theoretical ingestion points for indirect prompt injection, they are essential to the intended business logic and do not present an actionable vulnerability in the skill's own code snippets.\n- [EXTERNAL_DOWNLOADS]: The instructions reference official, well-known applications and services (Shopify, WooCommerce, BigCommerce, QuickBooks, Xero). These are documented neutrally as configuration steps for legitimate integrations and do not involve untrusted remote code execution.

b2b-commerce