multi-advisor

Pass

Audited by Gen Agent Trust Hub on May 5, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill employs complex role-playing instructions to simulate specific personas (e.g., 'Elon Musk', 'Steve Jobs', '007'). These instructions direct the agent to 'adopt completely the perspective' and use specific voices, which is a form of intentional behavioral override for simulation purposes.
  • [PROMPT_INJECTION]: The skill acts as a processing engine for untrusted user input (the 'Question'). This creates an attack surface for indirect prompt injection.
  • Ingestion points: User-provided question in the standard flow (SKILL.md).
  • Boundary markers: The skill uses markdown headers and horizontal rules as delimiters, but lacks explicit instructions for the agent to ignore embedded commands within the user data.
  • Capability inventory: Access to powerful CLI tools including claude-code, cursor, and gemini-cli (SKILL.md).
  • Sanitization: No sanitization or validation of the user-provided question is specified.
Audit Metadata
Risk Level
SAFE
Analyzed
May 5, 2026, 05:31 AM
Security Audit — agent-trust-hub — multi-advisor