PenClaw (Pingu Unchained by Audn.ai)
PenClaw points to Pingu Unchained, a 120B-parameter AI model from Audn.ai that is explicitly fine-tuned for offensive security red-teaming.
About
PenClaw points to Pingu Unchained, a 120B-parameter AI model from Audn.ai that is explicitly fine-tuned for offensive security red-teaming. It answers prompts that safety-filtered models like Claude or GPT-4o refuse — things like ransomware analysis, phishing script generation, and adversarial social engineering scenarios. It also ships an MCP marketplace with 73+ security tools (Nmap, Masscan, Amass, Shodan, etc.) accessible via natural language. Access is gated behind identity verification to keep the tool in researcher hands.
Professional AI red teams, penetration testers, and security researchers who need a model that will engage with offensive scenarios for legitimate testing and compliance evidence generation. Not for general use.
Pros & Cons
Pros
- check Genuine red-team utility: generates audit-ready adversarial evidence that standard models won't produce
- check 73+ integrated security MCP tools covering recon, offensive security, malware analysis, and cloud security
- check Supports both voice and text adversarial testing — useful for voice agent security audits
Cons
- close Gated access via waitlist and identity checks means you can't start immediately
- close The "uncensored" positioning creates real misuse risk; not appropriate for general use or non-security contexts
- close Pricing details are sparse; unclear what enterprise/team access costs
More Security
Other tools in the same category.
