
PyRIT
Open-source framework for red-teaming generative AI systems: automate attack prompts, score model responses, and audit behavior to identify security…

Open-source framework for red-teaming generative AI systems: automate attack prompts, score model responses, and audit behavior to identify security…

Inspect, debug, and visually test Model Context Protocol (MCP) servers from a web UI, CLI, or TUI, with tool/resource exploration, request logging,…

Unique tool for fingerprinting Large Language Models based on their tokenizers and behavior.

Security control plane for LLM agents: allowlists, owner kill switch, PIN sessions, rate limits, prompt-injection detection, and output scrubbing to…

FastGPT Python sandbox escape chain audit tool (CVE-2026-32128 related, v4.14.8 inspect chain)

Proof-of-concept demonstrating cross-channel trust fragmentation attacks on MCP-based AI coding assistants, splitting malicious instructions across…

Open-source AI-powered Security Operations Center — alert fusion, purple-team drills, agent-assisted triage, MITRE ATT&CK investigation.…

A font-based deception tool for red teaming, security research, and whatever else.

Runtime security gateway for AI agents: cryptographically attests tool calls, enforces policies, sandboxes execution, and logs tamper-evident audit…

AI-driven pentest harness with black-box, white-box, grey-box, host/cloud, and LLM red-team modes; validates findings with cross-model voting and…

A productionized greedy coordinate gradient (GCG) attack tool for large language models (LLMs)

Proof-of-concept tool for generating adversarial perturbations to exploit weaknesses in deep learning models, based on DEF CON 25 presentation.

Document ranking tool using pairwise LLM comparisons, designed for applications like vulnerability fix identification in code diffs.

Azure Sentinel detection lab for MCP attack patterns, providing 5 analytics rules and 7 KQL hunting queries against SSRF token theft, tool poisoning,…

Security benchmark for evaluating OpenClaw agents against adversarial execution contexts including poisoned files, injected skills, misleading tool…

Protocol-agnostic security kernel for AI tool execution. Enforces zero-trust, sandboxed isolation, policy-driven control, and full auditability…

Python library for local LLM-powered security analysis with Ghidra binary analysis, C/C++ vulnerability scanning, and MCP tool integration for…

Execution-Layer Security (ELS) for AI agents — policy-enforced shell with audit.