security-audit-methodology

Pass

Audited by Gen Agent Trust Hub on Jul 15, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [SAFE]: The skill consists exclusively of Markdown documentation providing guidance on threat modeling, vulnerability classification, and architecture review. No executable code, shell commands, or external dependencies were identified.
  • [NO_CODE]: The skill contains zero executable files or scripts. All reference materials are static documentation intended to guide the agent's reasoning process during an audit.
  • [DATA_EXFILTRATION]: While the reference materials describe common targets for data exfiltration (e.g., secrets in .env files or SSH keys) to help an auditor identify risks in a target system, the skill itself does not contain instructions or tools to access or exfiltrate such data from the environment where the agent is running.
  • [PROMPT_INJECTION]: The instructions establish a professional persona for security auditing and define a safety boundary requiring explicit authorization before any testing. It does not contain patterns attempting to override agent safety guidelines or bypass system constraints.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 15, 2026, 03:02 AM
Security Audit — agent-trust-hub — security-audit-methodology