competitor-profiling
Pass
Audited by Gen Agent Trust Hub on Aug 22, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns detected. The skill performs legitimate marketing research tasks using specialized external tools.- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external competitor websites via scraping tools (Firecrawl). Ingestion points: Firecrawl scrape and search functions in SKILL.md. Boundary markers: The synthesis process is separate from raw data storage, although no explicit security delimiters are used. Capability inventory: Limited to local file creation for profiles and raw data. Sanitization: None explicitly defined. The risk is minimized by the structured template-based output.- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill accesses project-specific context (e.g., product-marketing.md) but does not target sensitive system files or credentials. Network activity is confined to well-known research and scraping services (Firecrawl, DataForSEO) for gathering public market data.
Audit Metadata