reverse-engineer-any-store
Pass
Audited by Gen Agent Trust Hub on Nov 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from sources including the Meta Ad Library, competitor storefronts, and public buyer reviews. This activity creates a surface where malicious instructions embedded in third-party content could attempt to influence the agent's behavior.
- Ingestion points: Ad descriptions, product detail pages (PDPs), cart/checkout flows, and buyer feedback sequences.
- Boundary markers: The instructions recommend separating observations from inferences and citing sources but do not define specific technical context boundaries or delimiters.
- Capability inventory: The skill uses browser tools to navigate websites, inspect document structure, and simulate cart interactions.
- Sanitization: The instructions emphasize distinguishing between direct evidence and third-party reports, and explicitly warn against submitting private data to external forms.
- [EXTERNAL_DOWNLOADS]: The skill documentation includes links to external resources for extended functionality, including references to GitHub repositories from well-known services and the author's official catalog.
Audit Metadata