Havuz

给你的AI订一份晨报：常驻agent的信息摄入与跨context认知循环（思路篇，无代码）

Closure（可露希尔）— 开源、本地优先的 AI 小说创作 IDE。Open-source, local-first AI novel-writing IDE. Fork of OrisonSpace.

The ops layer for running Claude Code and Codex all day on a Mac: subscription-first routing with overflow to GLM/OpenRouter/Bedrock, pausel

Lemmalog, LLM agent’ların belleğini Datalog kurallarıyla temsil edip kaynak izlenebilir ve artımlı biçimde sorgulayan bir Rust motorudur.

Neden yükseliyor · Agent memory ve MCP konularına yönelik ilgiyle örtüşüyor; 31 günlük repo için 324 yıldızın yanı sıra son 28 günde 27 commit ve 4 farklı yazar var. Bu hareketlilik erken geliştirme temposuna işaret ediyor, ancak yükselişin nedenini doğrudan kanıtlamıyor.

opengrok, mevcut Grok Bot kurulumunda agent’lere farklı LLM’ler bağlamak ve model ayarlarını yönetmek için bir config sidecar sunuyor.

Neden yükseliyor · 31 günlük yaşına karşılık 464 yıldız ve son 28 günde 28 commit dikkat çekiyor. README’deki model picker, sağlayıcıya özel ayarlar ve Grok Bot içi panel, dar bir kullanım alanında somut bir ürün deneyimi sunuyor; ancak son 28 gündeki 2 farklı yazar, geliştirme tabanının sınırlı olduğuna işaret ediyor.

Local-first persistent memory for OpenAI Codex: governed recall, progressive disclosure, no hosted vector database.

QuantBot（AI 量化机器人）的全量源码。基于 **Go+React + TypeScript + Vite** 构建，通过 **Wails** 与 Go 后端通信，提供 A 股量化交易的图形界面：六维判势、多因子选股、AI 智能体决策、策略回测、实时活动、审计日志等。

Learn AI agents from first principles to production. Zero mandatory dependencies, pure standard Python 3.11+, and no magic frameworks.

TTPO: Test-Time Policy Optimization

把小说角色变成"活人"：每个角色是一个独立模拟体，只知道他该知道的，怕他该怕的，说他自己的话。 剧情线是引力场，不是铁轨——角色自由行动，世界只用"世界内的手段"把故事拉向既定节点。

The Conifer SDK — one gateway for cloud, local, and BYOK inference, with exact per-call cost receipts. TypeScript + Python + MCP.

曾仕强 AI Agent Skill，基于曾仕强书籍，电视节目及公开演讲等材料构建的大模型知识库 Skill 资源库。完整收录中国式管理全集 23 册、易经、周易、道德经、论语、财神文化、大易管理、三国智慧、家风家庭教育等人情世故与国学管理文本，全部整理为标准化 Markdown

PromptForge scores LLM prompts across seven quality dimensions, then rewrites weak prompts into clear, intent-preserving instructions

Dependency-free Telegram interaction core for model-driven bots: ordered bubbles, reactions, regular stickers, and one-shot choices over hos

Русский Humanizer: Agent Skill и линтер читаемости для LLM-агентов

Playwright, but built for AI agents: the browser API models already know, token-efficient a11y snapshots, and scroll-driven collection over

Training-data memorization auditor for fine-tuned LLMs — Trainer/TRL plugin, canary MIA + regurgitation audit, Apache-2.0

MoE expert compression toolkit — prune, merge, and forge specialist models for local inference.

halogen-flash-server, Qwen3.8-Flash-Next modelini AMD Strix Halo (gfx1151) donanımında çalıştırmak için özelleştirilmiş bir inference server'dır.

Neden yükseliyor · 32 günlük repo için 722 yıldız ve son 28 günde 99 commit dikkat çekiyor. README, belirli donanım üzerindeki ölçümleri ve ölçüm koşullarını açıklıyor; ancak performans iddiaları proje tarafından sunulmuş ve bağımsız doğrulama bilgisi burada yok. 99 commit'in tek yazardan gelmesi geliştirme hızını gösterirken bakım kapasitesine dair kanıtı sınırlıyor.

Pentest Harness, yetkili güvenlik testleri için farklı LLM sağlayıcılarını ve pentest araçlarını yerel bir arayüzde birleştiren self-hosted bir agent çalışma alanıdır.

Neden yükseliyor · 33 günlük bir repoda 398 yıldız dikkat çekiyor; ancak README dışındaki benimseme verisi bilinmiyor ve son 28 günde 2 commit ile 2 farklı yazar görülmesi geliştirme temposu hakkında güçlü kanıt sunmuyor.

Halo is an open-source framework built by White Circle for training large language and multimodal models

CDAF (Cached Descriptive Asset Files) - open sidecar format for video so AI agents stop re-analyzing the same footage. Spec, CLI, agent skil

The fastest way to run Qwen3.8 27B on Strix Halo (gfx1151)

An open-source take on Grok Bot/Muse. Persistent AI teammates with their own cloud computers, powered by Runta.

Human-feedback and data-governance infrastructure for RSI research: annotation, review, provenance and frozen dataset candidates.

The open startup repo — everything a founder needs to start and run a company in the agentic era: the comprehensive founder processes from a

Agent workflow for deadline-bound execution with evidence gates

CodeMonk 🥋 —>An open-source AI powered code intelligence platform that helps developers understand, explore, and navigate complex codebases

1-Click Grok Bot 8 Man Crypto Trading Desk as seen on X

The Military-Grade Memory Hygiene, private, offline-first prompt manager for AI Agent power users. Features AES-256 encrypted local storage,

基于多智能体架构与自主决策机制，配合大语言模型的理解能力，输入自然语言指令，AI 自动完成从目标侦察到漏洞验证的完整渗透测试流程。

Open Memory Protocol (OMP): an open protocol for moving and sharing AI-agent memory. Hosted at the AI Disclosures Project, co-convened with

Self-hosted AI agents with persistent memory, browser and computer access, and scheduled workflows.

Skills to bootstrap and expand analytics ontologies from schema, dbt, dashboards and docs. Evidence-backed context, user-defined domains and

Pyramis-L3D: hierarchical sparse attention research prototype (code only; weights on Hugging Face)

Full-text financial reports (US, China, Japan, Korea, Taiwan) as clean Markdown — REST API + MCP server for LLMs, RAG and quant

Your AI forgets. A wiki doesn't. Turn chat-pasted text/links into a verified, translated, searchable wiki knowledge base — verify links, fil

Unofficial ACP adapter for Meta's Muse Code — use Muse in Zed, VS Code, and other Agent Client Protocol editors with streaming, tools, sessi

A hosted AI coworker in the channels where your team already talks — open source, and yours to run.

A Rust crate to connect multiple harnesses under one, simple, up-to-date interface.

Wemux - self-hostable AI agent collaboration platform (Apache-2.0, community edition)

Your pocket agent. Local-first AI agents on iOS and Android — real workspaces, tool execution with approvals, and your choice of model (DSH

对抗模型输出不良表达习惯的 Agent Skill

Privacy-focused AI conversation platform for chat, coding, research, and documents.

Build & Auto-Optimize LLM Apps in 2026: The Ultimate AdalFlow Guide

Ultimate LLM Gym Trainer Guide 2026: Reinforcement Learning Fine-Tuning Framework

Master LLM Fine-Tuning Trainer 2026: Advanced NLP Optimization Toolkit

Bilingual AI study coach that answers only from your own course material, with citations

Give your CLAUDE.md / AGENTS.md a checkup — audit size vitals, dead references, drifted claims, and backtest every rule against your own ses

A 21-day hands-on journey to mastering AI for DevOps — covering LLMs, GenAI, CI/CD, Kubernetes, cloud, automation, and AI-powered DevOps pro