deep-think

Pass

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: LOWSAFENO_CODE
Full Analysis
  • [Indirect Prompt Injection] (LOW): The skill is designed to process untrusted user queries for analysis. While it lacks explicit boundary markers or sanitization, the skill possesses no capabilities for command execution, file system access, or network communication, restricting potential impact to reasoning bias.
  • [No Code] (SAFE): The skill consists entirely of markdown instructions and examples, with no scripts or executable content.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 16, 2026, 05:26 AM