paul-graham

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • NO_CODE (SAFE): All nine files in the skill are markdown documents (.md). There are no scripts (Python, JS, etc.), binaries, or executables included in the package.
  • EXTERNAL_DOWNLOADS (SAFE): The skill does not contain any configuration files (like package.json or requirements.txt) that would trigger external package installations. No remote code patterns were found.
  • DATA_EXFILTRATION (SAFE): There are no network-related commands (curl, wget) or file system operations that attempt to access or transmit sensitive information.
  • PROMPT_INJECTION (SAFE): The instructions in SKILL.md focus on establishing a persona ('direct', 'opinionated') for providing advice. They do not include directives to ignore safety filters, reveal system prompts, or bypass AI constraints.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:15 PM