ulta-research
Pass
Audited by Gen Agent Trust Hub on Oct 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external product data, customer reviews, and Q&A content from the Crawlora API, which constitutes an indirect injection surface.
- Ingestion points: Data is retrieved from
https://api.crawlora.net/api/v1via thescripts/crawlora.shscript (e.g., inSKILL.md). - Boundary markers: While the skill instructions specify returning "normalized JSON," they do not mandate the use of explicit delimiters (like XML tags or triple backticks) to isolate external content in the final response.
- Capability inventory: The skill is restricted to making specific GET requests to the vendor's API. It has no capabilities for file-system modification, privilege escalation, or arbitrary command execution.
- Sanitization: The
scripts/crawlora.shhelper script includes extensive sanitization, including alphanumeric validation of the API key, HTTP method restriction (GET only), path whitelisting via regex, and mandatory URL encoding for all query parameters.
Audit Metadata