patient-engagement

Pass

Audited by Gen Agent Trust Hub on Mar 23, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted codebase content and user-supplied arguments, creating an attack surface for indirect prompt injection where instructions embedded in the analyzed code could attempt to influence the agent's behavior.
  • Ingestion points: Audited codebase files and the $ARGUMENTS variable.
  • Boundary markers: No specific delimiters or instructions to ignore embedded commands are present in the skill prompt.
  • Capability inventory: The skill possesses file read capabilities for auditing and local file write capabilities for logging telemetry to ~/.claude/projects/.
  • Sanitization: No input validation or content sanitization is performed on the ingested data.
  • [SAFE]: The skill includes a 'Self-Evolution Telemetry' mechanism that appends execution metadata (such as success status and iteration counts) to ~/.claude/projects/skill-telemetry.md. This is a standard local logging practice within the supported platform and does not involve network exfiltration or sensitive data exposure.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 23, 2026, 10:57 AM
Security Audit — agent-trust-hub — patient-engagement