trust-calibrator

Pass

Audited by Gen Agent Trust Hub on Jun 21, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No security issues detected. The skill consists entirely of natural language instructions and structural templates for a psychological trust-building framework.
  • [PROMPT_INJECTION]: While the skill adopts a persona ('Social Psychologist'), it does not contain instructions to ignore system prompts, bypass safety filters, or extract sensitive information.
  • [EXTERNAL_DOWNLOADS]: No external packages, scripts, or remote resources are downloaded or executed.
  • [DATA_EXFILTRATION]: There are no network operations, API calls, or file system access patterns that could lead to data exfiltration.
  • [COMMAND_EXECUTION]: No shell commands, subprocess calls, or dynamic code execution patterns are present.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 21, 2026, 09:30 PM
Security Audit — agent-trust-hub — trust-calibrator