Skills
Navigate
skills/brevdev/nemoclaw-demos/healthcare-payer-audit/Gen Agent Trust Hub

healthcare-payer-audit

Pass

Audited by Gen Agent Trust Hub on Jul 22, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Executes a local script at /sandbox/.openclaw/workspace/scripts/care_backlog_analyzer.py to analyze backlog data.
  • [PROMPT_INJECTION]: The skill processes untrusted data from local JSON files, creating a surface for indirect prompt injection. Ingestion points: /sandbox/.openclaw/workspace/data/payer_rules.json and escalation_rules.json. Boundary markers: None identified. Capability inventory: Execution of a local Python script via subprocess. Sanitization: No validation or sanitization of the JSON content is documented.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 22, 2026, 08:28 AM
Security Audit — agent-trust-hub — healthcare-payer-audit