aicodeking / leanrea
One 8 KB skill that replaces REA (Reverse Engineer Anything): plain CLI tools, no MCP. Benchmarked: 38% faster, 35% cheaper.
aicodeking / leanrea
MemPalace / mempalace
The best-benchmarked open-source AI memory system. And it's free.
tirth8205 / code-review-graph
Local-first code intelligence graph for MCP and CLI. Builds a persistent map of your codebase so AI coding tools read only what matters, with benchmarked context reductions on reviews and large-repo workflows.
AGIHunt / blurt
Show it. Say it. Your AI gets it. Record your screen and talk — your coding agent turns it into bugs, ideas and to-dos, with the exact frames marked. 口喷鸡：边看边喷，AI 全懂。Agent Skill + macOS menu-bar app for Claude Code / Codex.
bonnhatnguyen / dola-render-gateway
Dola Studio — AI Video Generation & Render Gateway (Seedance 2.5/2.0, 100% Unwatermarked Extractor, Browser Pool)
jack66g / spatial-memory-engine
A next-gen long-term memory engine for AI apps: self-organizing spatial regions, memory dynamics (Ebbinghaus reinforcement / decay / consolidation), and explainable 8-signal retrieval. Runs offline with zero config; works with any OpenAI-compatible LLM/embedding, cloud or local. Chinese-friendly. Benchmarked head-to-head vs Mem0 and bare RAG
0xBakeer / Qwen3.8-27B-FP8-on-a-single-DGX-Spark
Serving Qwen3.8-27B-FP8 on a single DGX Spark (GB10): 7.88 to 58.5 tok/s single-stream from decode strategy alone, weights untouched. Speculative decoding and prefix caching benchmarked, plus DFlash 2 — the only Qwen3.8-27B build that can serve it under vLLM.
tonyd2wild / DS4-H3-Video-Gen-Factory
Run DeepSeek-V4-Flash at full 1M context AND two MiniMax H3 video instances on the same two DGX Sparks. Benchmarked C1-C6.