personize-diagnostics

Pass

Audited by Gen Agent Trust Hub on Mar 14, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill does not contain any malicious code, obfuscation, or unauthorized data exfiltration patterns.
  • [DATA_EXPOSURE]: The skill correctly uses environment variables (e.g., PERSONIZE_SECRET_KEY, TRIGGER_SECRET_KEY) to manage sensitive credentials, which is the recommended security practice for SDK integrations.
  • [EXTERNAL_DOWNLOADS]: The skill references the official @personize/sdk and @trigger.dev/sdk. These are trusted libraries from the primary vendor and a well-known automation service, respectively.
  • [INDIRECT_PROMPT_INJECTION]: The skill involves reading data from external memory stores to verify system health. While processing untrusted external data is an inherent surface for indirect prompt injection, this skill operates as a diagnostic tool where inspecting such data is the primary intended function. There is no evidence of unsafe command execution or privilege escalation tied to this data processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 14, 2026, 07:00 AM