skills/neurofoo/agent-skills/redteam/Gen Agent Trust Hub

redteam

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [PROMPT_INJECTION] (SAFE): While the skill uses adversarial framing ("Think like an attacker", "break this"), these instructions are scoped to the legitimate task of analyzing a user-provided plan for weaknesses. It does not contain instructions to bypass AI safety filters or disregard system prompts.
  • [NO_CODE] (SAFE): The skill does not include any Python scripts, Node.js code, or shell commands. It is purely a structured text template.
  • [DATA_EXFILTRATION] (SAFE): There are no mechanisms for file access or network communication, preventing any risk of data being sent to external servers.
  • [REMOTE_CODE_EXECUTION] (SAFE): No external dependencies, package managers, or remote script download patterns were identified.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:04 PM