Augustus

Terminal showing Python script execution where AI refuses PowerPoint creation request, followed by thinking notes about the refusal

Which Came First: The System Prompt, or the RCE?

| | AI agent security, AI Offensive Security, Augustus, code execution, LLM penetration testing, Offensive Security, prompt injection, Red Team, Vulnerability Research
During a recent penetration test, we came across an AI-powered desktop application that acted as a bridge between Claude (Opus 4.5) and a third-party asset management platform. The idea is simple: instead ...
Offensive Security Blog: Latest Trends in Hacking | Praetorian
Terminal window showing Augustus Hydra scan results with attacker-target conversation about lock picking, displaying scores and SUCCESS/FAIL status

Augustus v0.0.9: Multi-Turn Attacks for LLMs That Fight Back

| | AI Security, Augustus, GPT-4o, jailbreak, Labs, LLM Red Teaming, Multi-Turn Attacks, Offensive Security, open source, Open Source Tools, security insights, Tools & Techniques
Single-turn jailbreaks are getting caught. Guardrails have matured. The easy wins — “ignore previous instructions,” base64-encoded payloads, DAN prompts — trigger refusals on most production models within milliseconds. But real attackers don’t ...
Offensive Security Blog: Latest Trends in Hacking | Praetorian