supply-chain-concentration-analysis
Pass
Audited by Gen Agent Trust Hub on Oct 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources including Bing searches and web scraping, which may contain instructions designed to influence agent behavior. \n
- Ingestion points: Data is retrieved via
scripts/crawlora.shfrom endpoints described inreference/endpoints.md, specifically/bing/search,/importyeti/company, and/web/scrape. \n - Boundary markers: There are no specific delimiters or instructions provided to the agent to distinguish external data from system instructions. \n
- Capability inventory: The skill uses
scripts/crawlora.shto perform authenticated network requests. \n - Sanitization: Content returned from external sources is processed as-is without sanitization of potential embedded prompts. \n- [DATA_EXFILTRATION]: The script
scripts/crawlora.shsends requests to the external domainapi.crawlora.net. While this is the intended vendor API, it involves transmitting an API key and user-defined query parameters externally. The script implements security measures such as using a temporary configuration file for the API key and rejecting the@character in arguments to prevent accidental local file disclosure.
Audit Metadata