competitor-data-verifier

Pass

Audited by Gen Agent Trust Hub on Apr 8, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill is designed to gather public information from competitor domains and documentation subdomains, which aligns with its intended purpose.
  • [SAFE]: A strict no-auto-modify policy is enforced, ensuring that all data updates are presented as recommendations in a report for human validation.
  • [SAFE]: Analysis of the reference files shows the use of zero-width space characters (U+200B) in markdown templates as a benign escaping mechanism for documentation.
  • [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection through the ingestion of external web content. Ingestion points: WebFetch and WebSearch results from competitor websites; Boundary markers: absent; Capability inventory: Read and Glob operations (no write or shell execution capabilities); Sanitization: none. This surface is assessed as safe because the agent's actions are restricted to comparative analysis and report generation with human oversight.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 8, 2026, 02:57 AM