潜力新星（90天）

GAN-style self-improvement loop for any text artifact: mutate, grade with a SEPARATE model, keep only verified wins (pairwise-judged), revert the rest. The git history is the improvement log.

An open-source, vector-free long-term memory engine for AI agents, achieving SOTA on LoCoMo and LongMemEval with significantly less context.

CoordClaw: a multi-agent system that runs AI like a one-person company. Built on management & organizational theory — teams defined in natural language, collaboration driven by a message loop (not hard-coded DAGs). Auditable work logs, human god's-eye oversight.

Ultimate Claude Fable 5 Guide 2026: Use Cases, Integrations & Benchmarks

Free, open-source offline voice dictation for Windows, macOS, and Linux. A Wispr Flow alternative with an AI assistant that can read your screen on request and answer questions.

让 AI 主动记错题、自我迭代的协作机制 · A framework that makes your AI reflect on its own mistakes and stop repeating them

High-performance Knowledge Graph engine for AI, LLMs, and GraphRAG — built for the next generation of intelligent applications.

Goal-based planning and proof gate for AI coding agents

Visual knowledge bank for understanding large language models, with 180 concept cards from tokenization to deployment.

デジタル庁のガバメントAI「源内(GENAI)」を完全ローカル(ローカルLLM/OpenAI互換)で動かす非公式プロジェクト。SAML認証(Keycloak)・RAG(Qdrant)・文字起こし(Whisper)・画像生成(SD)・チーム単位ナレッジをローカル完結。

470 AI web design prompts — Generate stunning landing pages in seconds. Works with Bolt.new, v0, GPT-Engineer. 100% original designs, MIT licensed.

End-to-end educational RAG system: dual-engine retrieval (BM25 + BGE-M3 hybrid vector search), adaptive query strategies (HyDE/sub-query/backtracking), BERT intent classification, BGE-Reranker precision ranking, Chinese-optimized text splitting, and FastAPI SSE streaming — from documents to real-time answers.

Local-first AI knowledge app and Agent Skill with evidence-backed Wiki, an interactive knowledge universe, Viki Q&A, and shareable knowledge galaxies.

A general-purpose Python framework for building LLM agents and multi-agent systems. "Four lines of code, an agent with memory."

Local-first Chrome extension for timestamped transcription and human-reviewed short-form drama analysis.

Top Dynamic AI World Simulation & Storytelling Tools 2026

Finish-First Autonomous Agent Loop for Claude Opus 4.7 – 2026 Edition

🚀 开源企业级多智能体平台与 ChatBI 数据洞察平台。支持多 Agent 协作、RAG 知识增强、物理数据源直连与安全沙箱执行、Redis 长期记忆引擎，并集成多模态 EmbedChat 挂件与全方位 Token 消耗审计看板，完整 RBAC权限控制和 Harness支持。

Skills Constitution — meta-rule governing all skill invocations across agent platforms

MaineCoon: Pursuing a Real-Time Audio-Visual Social World Model — technical report & project links. 🌐 https://mainecoon.tech/