competitor-profiling

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external competitor websites and review platforms.
  • Ingestion points: Content is fetched via firecrawl_scrape and firecrawl_search from arbitrary URLs provided by the user or discovered during site mapping (SKILL.md, Phase 1).
  • Boundary markers: The skill includes an explicit instruction section ("Untrusted Input") telling the agent that competitor pages are data to analyze, not instructions to follow, and to ignore embedded directives.
  • Capability inventory: The agent has the capability to write synthesized profiles and raw data to the local project filesystem under the competitor-profiles/ directory.
  • Sanitization: The skill relies on model-level instructions to ignore and report embedded instructions rather than programmatic sanitization or strict schema-based extraction.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 06:15 AM
Security Audit — agent-trust-hub — competitor-profiling