competitor-discovery
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from external competitor websites during the profiling phase, which could allow a website to influence the agent's behavior.
- Ingestion points: In Step 4, the skill fetches homepages and pricing pages of identified candidates to extract one-liners and audience signals.
- Boundary markers: The instructions lack specific boundary markers or "ignore instructions" warnings when processing the fetched site content.
- Capability inventory: The skill utilizes web search and URL fetching tools to retrieve data from the public internet.
- Sanitization: There is no evidence of sanitization or filtering applied to the retrieved HTML/text content before the agent analyzes it.
Audit Metadata