competitor-profiling
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data by scraping external websites using the
firecrawl_scrapetool. This content is then synthesized into profiles, creating a potential surface for indirect prompt injection attacks where a competitor's website could contain malicious instructions intended to influence the agent's behavior or output. - Ingestion points: External webpage content and metadata retrieved via the
firecrawl_scrapeandfirecrawl_maptools as described in Phase 1 ofSKILL.md. - Boundary markers: The instructions lack explicit guidelines for using delimiters or boundary markers to isolate untrusted scraped content from agent instructions during the synthesis phase.
- Capability inventory: The skill has capabilities for network communication (via Firecrawl and DataForSEO MCP tools) and local file system writes (saving data to the
competitor-profiles/directory). - Sanitization: There are no defined steps for sanitizing, filtering, or validating the scraped markdown content before it is processed by the agent to create the final profile documents.
Audit Metadata