kelly-insurance-intel

Pass

Audited by Gen Agent Trust Hub on Aug 25, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill correctly implements a gated review process where all externally-sourced data, including signals, actions, and client drafts, is stored in Busabase for manual inspection. No external side effects occur without human intervention and approval through the AirApp UI.\n- [INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it processes data from external industry news and regulator sources, which is mitigated through clear boundaries.\n
  • Ingestion points: The agent browses public insurance news, regulator portals, and market signals as part of the daily intelligence collection workflow.\n
  • Boundary markers: All signals and actions are held in a 'needs review' state. The SKILL.md provides explicit instructions to block automated advice and outbound messages.\n
  • Capability inventory: The skill uses the busabase-sdk for gated storage and administrative scripts for status updates.\n
  • Sanitization: The AirApp frontend and model logic employ HTML escaping and strictly typed JSON parsing to safely handle untrusted input strings.\n- [COMMAND_EXECUTION]: The skill provides administrative scripts for environment setup and decision tracking. These scripts are benign and follow safety practices, such as using environment variables for authentication and defaulting to dry-run modes to ensure the operator can review changes before applying them.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 25, 2026, 10:36 AM
Security Audit — agent-trust-hub — kelly-insurance-intel