persona-team-lead

Pass

Audited by Gen Agent Trust Hub on Jul 15, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill defines persona-based workflows using gws binary commands and does not exhibit any signs of command injection, credential harvesting, or unauthorized data exfiltration.
  • [PROMPT_INJECTION]: Processing data from Gmail and Google Chat presents a surface for indirect prompt injection. 1. Ingestion points: Gmail and Chat via gws workflows in SKILL.md. 2. Boundary markers: Absent. 3. Capability inventory: Workspace management via the gws binary. 4. Sanitization: Instructions explicitly advise using the --sanitize flag to handle sensitive data.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 15, 2026, 01:27 PM
Security Audit — agent-trust-hub — persona-team-lead