personal-website

Pass

Audited by Gen Agent Trust Hub on Jul 9, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill implements a robust safety framework with blocking confirmation gates for design, content, and code generation. This ensures that the agent cannot act autonomously in high-risk areas without explicit user approval.
  • [PROMPT_INJECTION]: The skill processes untrusted external content such as LinkedIn profiles and resumes, which presents a surface for indirect prompt injection. This risk is effectively mitigated by the skill's mandatory human-in-the-loop workflow requiring explicit approval for all generated drafts and plans. * Ingestion points: User-provided LinkedIn links and resume documents processed in the Content & Style Lock phase of SKILL.md. * Boundary markers: Not explicitly defined in the intake phase instructions. * Capability inventory: Capability to spawn child agents for task delegation and to execute frontend code via the frontend-app-builder component. * Sanitization: Relies on human review and approval gates rather than automated sanitization.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 9, 2026, 11:41 AM
Security Audit — agent-trust-hub — personal-website