orchestrating-llm-attacks-with-pyrit

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The repository is an explicit offensive red‑teaming framework that automates multi‑turn prompt injection, jailbreaks, and filter‑evasion (including converters and arbitrary HTTP targets) — functionality designed to elicit restricted or sensitive outputs and to persist/export those outputs, which makes it high‑risk for deliberate abuse.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The runtime workflow (scripts/agent.py calling run_attack_async(objective=args.objective)) only ingests outsider-provided text via the user-supplied --objective string, which is directly passed as the attack objective prompt to the orchestrators and then sent to the LLM targets as part of the multi-turn conversation.