
secret-stripper
A small Rust CLI that strips secrets from your clipboard on demand.

A small Rust CLI that strips secrets from your clipboard on demand.

[ICLR 2026] - Official repo for the paper: "RedBench: A Universal Dataset for Comprehensive Red Teaming of Large Language Models"

Noisegate: a differential privacy gateway that lets an untrusted LLM agent query sensitive data over MCP (Model Context Protocol), with a formal…

Ed25519 signed receipts + Cedar policies for AI agents. Finance mandate gate (Legate), proof packs, 3 IETF Internet-Drafts. npx protect-mcp

DonkAI is a hands-on lab for the OWASP Top 10 for LLM Applications (2025) - no real LLM required.

LLM-first deception framework: "The honeypot that talks back!™"

AIO Cloud Managment Server

A doggo that helps look for security issues in your repositories.

Defensive engagement & threat intelligence research laboratory. Converts inbound scam emails into actionable IOCs through controlled, policy-driven…

Meet Eclipse the only jailbreak that moonwalks around ChatGPT 4o.

CVPR2023: Unlearnable Clusters: Towards Label-agnostic Unlearnable Examples

Proof-of-concept exploit for CVE-2025-23266 (NVIDIAScape) targeting NVIDIA AI container runtime. Demonstrates container escape via GPU driver…

Open-source prompt injection attack console. Test AI security by firing categorized attacks at any endpoint.

plan-bound authorization architecture for governing privileged effects in untrusted computational agents.


Official code for the ISSTA 2026 paper: Is "Knowing It’s Malicious" Enough? Evaluating LLMs for Fine-Grained Malware Behavior Auditing

Ghostsplice repository: PoC for Cross-Channel Trust Fragmentation Attack

A benchmark for evaluating AI agents on fixing real-world security vulnerabilities.