guru-paul-graham

Pass

Audited by Gen Agent Trust Hub on May 1, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill's primary function is to provide startup advice by referencing a fixed set of 14 local Markdown files. These files are standard copies of public essays and do not contain executable code, obfuscated content, or malicious instructions.
  • [SAFE]: Data handling is restricted to reading from the skill's reference directory and writing a summary of user interactions to a specific knowledge path (knowledge/yc-startup-school/guru-pg-workbook.md). No unauthorized file access or network exfiltration patterns were observed.
  • [SAFE]: The prompt instructions define a specific expert persona ('Guru Paul Graham') and instructional steps. These do not attempt to bypass safety filters or override the agent's core operational guidelines.
  • [SAFE]: The skill uses the standard AskUserQuestion tool for its interactive diagnostic component, following a logical and safe flow of user-provided information for the purpose of generating advice.
Audit Metadata
Risk Level
SAFE
Analyzed
May 1, 2026, 10:27 PM
Security Audit — agent-trust-hub — guru-paul-graham