moat-analysis

Pass

Audited by Gen Agent Trust Hub on Jul 12, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill involves processing user-provided business ideas and performing external web research, creating a surface for indirect prompt injection where malicious instructions could be embedded in researched content. This risk is inherent to the research-focused nature of the skill.
  • Ingestion points: User input in Step 0 and web research results in Step 1.
  • Boundary markers: No specific boundary markers or instructions to ignore embedded commands are present in this file.
  • Capability inventory: The skill uses research capabilities but does not exhibit dangerous operations like shell execution, network exfiltration, or local file writing.
  • Sanitization: There is no evidence of data sanitization or input validation for external content.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 12, 2026, 12:36 AM
Security Audit — agent-trust-hub — moat-analysis