taxue-roundtable

Pass

Audited by Gen Agent Trust Hub on Mar 29, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill uses structured role-playing prompts for historical figures (e.g., Drucker, Munger). There are no instructions identified that attempt to override system safety guidelines or bypass AI constraints.
  • [DATA_EXFILTRATION]: No network access or commands for reading sensitive local configuration files (like .ssh, .aws, or .env) were detected.
  • [COMMAND_EXECUTION]: The skill defines logic for routing tasks to other modules within its own ecosystem (e.g., /solve, /build, /learn), but it does not execute arbitrary shell commands or attempt privilege escalation.
  • [REMOTE_CODE_EXECUTION]: The skill does not download external scripts or packages, and it lacks dynamic execution functions like eval or exec.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 29, 2026, 11:44 PM