Starred repositories
Agent skill for hill-climbing local LLM serving speed (decode, speculative decoding, prefill, KV, startup, TTFT) on any hardware and quantization
Open source reimplementation of GTA San Andreas game engine in Unity
Tool for automatic PS5 executables porting to Linux and Windows
Modern coding harness, terminal UI and web-based
The website QA tool for your coding agent. 350+ audit rules across SEO, performance, security, accessibility and agent experience.
Run Claude Code, Codex, Copilot and other agent CLIs in parallel worktrees, then check what each one did before you merge.
Universal provider proxy for OpenAI Codex & Claude Code — use any LLM (Claude, Gemini, Grok, DeepSeek, Ollama…) with Codex CLI, App, SDK, and Claude Code
Automatic, model-driven permission review for OpenCode
Dynamic context pruning plugin for OpenCode - intelligently manages conversation context to optimize token usage
Qwen3.8-Flash-Next (125B MoE) up to 1M context on one or two RTX cards (tuned on 3080 10gb). A measured, tuned and benchmarked fork of Strata.
GitHub (Primer) look-and-feel theme for Gitea 1.27.x and 28.x — light, dark, auto
DirectStorage for Windows is an API that allows game developers to unlock the full potential of high speed NVMe drives for loading game assets.
Repurpose Walmart Hanshow Stellar e-ink price tags with a Flipper Zero or Python. Requires ATC_TLSR_Paper firmware on the tags.
A tool to unlobotomize your NVIDIA card!
Aristo94 / EngramHalo.cpp
Forked from ggml-org/llama.cppllama.cpp fork tuned for Qwen 3.8 Flash-Next on AMD Strix Halo — RDNA3.5 kernel patches, true QSA sparse gather, working MTP, 27 GB engram table on SSD. 23.5 → 39 tok/s.
Fast, small, and fully autonomous AI personal assistant infrastructure, any OS, any platform — deploy anywhere, swap anything 🦀
Qwen3.5-2B with a SIMT GPU inside its weights, 45 compiled layers in front of Qwen’s own, using its Gated DeltaNet layers as registers and memory.
Custom tools to enhance your Open-Webui Experience! 🚀
Caveman Compression is a semantic compression method for LLM contexts. It removes predictable grammar while preserving the unpredictable, factual content that defines meaning.
ASD-STE100 Simplified Technical English rules, repurposed as a Claude Code skill for rewriting ambiguous agent-facing English.
A full retained fly-connectome simulation with experimental memory and guarded Coinbase AgentKit trading actions.
A 3D fruit fly living on your macOS desktop, driven by a live spiking simulation of the real FlyWire connectome
Unlock vGPU functionality for consumer grade GPUs.