Starred repositories
Intelligent Mixture-of-Models Router for Efficient Heterogeneous LLMs Inference
TUI tool for managing multiple Claude Code sessions in parallel using tmux
"Explaining Attention with Program Synthesis" | Work to reproduce analyzing and replacing attention heads from MIT LINGO '25 Summer Research Program
Kernel-level eBPF sandbox for securing LLM agent tool calls made through the Model Context Protocol (MCP)
Culture turns isolated stochastic agents into cooperative, inspectable, improvable artificial colleagues.
Algorithm powering the For You feed on X
Can Language Models Rebuild Programs From Scratch?
SEAL 911 is a project designed to give users, developers, and security researchers an accessible way to connect with a small group of highly trusted security professionals in case of emergency.
Fully automatic censorship removal for language models
OpenShell is the safe, private runtime for autonomous AI agents.
This repository contains the HTML source code discussed in my Medium article.
List of Permanent Free LLM API (API Keys)
Deep Learning Interpretability with Symbolic Regression
Secure, Fast, and Extensible Sandbox runtime for AI agents.
A small Recursive Language Model: let any LLM run code on its context instead of stuffing it into the prompt.
An official implementation of the paper: "Advancing Image Classification with Discrete Diffusion Classification Modeling".
OWASP Foundation web repository
A modular tool to search for known vulnerabilities, exploits and more across various data sources
Unified MCP server for querying OpenTelemetry traces across multiple backends (Jaeger, Tempo, Traceloop, etc.), enabling AI agents to analyze distributed traces for automated debugging and observab…
The minimalist Markdown viewer for busy terminals. Pure Bash. Zero dependencies. Works everywhere.
CUGA is an open-source generalist agent harness for the enterprise, supporting complex task execution on web and APIs, OpenAPI/MCP integrations, composable architecture, reasoning modes, and policy…
🎒 Token-Oriented Object Notation (TOON) – compact, human-readable serialization of JSON data for LLM prompts. TypeScript SDK, CLI, benchmarks.
Making open safety AI models accessible and beneficial to the safety community
A lightweight sandboxing tool for enforcing filesystem and network restrictions on arbitrary processes at the OS level, without requiring a container.
Google Research
AgentFlow: In-the-Flow Agentic System Optimization
Python tool for converting files and office documents to Markdown.
PairDrop: Transfer Files Cross-Platform. No Setup, No Signup.