seo-ecommerce
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes Python scripts with user-supplied inputs as arguments, such as
<url>or<keyword>. Without proper sanitization, this provides a surface for command injection.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external product pages and marketplace APIs.\n - Ingestion points: Page fetching and API results in
SKILL.md.\n - Boundary markers: None identified.\n
- Capability inventory: Local script execution and file writes to
.seo-cache/.\n - Sanitization: Not specified for external content.\n- [DATA_EXFILTRATION]: The skill reads and writes to a local cache directory (
.seo-cache/) containing sensitive domain and industry metadata, which could be targeted if file system access is compromised.
Audit Metadata