claims-workflow

Pass

Audited by Gen Agent Trust Hub on Mar 23, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill is designed to generate a comprehensive analysis report and save it to docs/claims-workflow-analysis.md. This is a standard output function for an analysis tool.- [DATA_EXPOSURE]: The skill includes a 'Self-Evolution Telemetry' phase that attempts to read from and write to ~/.claude/projects/skill-telemetry.md. This is used to log outcome metadata for skill improvement within the agent's execution environment and does not involve exfiltration to external domains.- [SAFE]: The skill explicitly includes safety constraints that prevent the agent from modifying claims data, accessing personally identifiable information (PII), or making legal/medical determinations.- [SAFE]: The discovery phase (Phase 1) uses standard project file detection (e.g., searching for pom.xml, package.json, requirements.txt) to identify the underlying technology stack without executing the files found.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 23, 2026, 10:56 AM
Security Audit — agent-trust-hub — claims-workflow