Havuz

50x your usage with cheap-ag. Opus/Astra plans, a dirt-cheap model writes every line, and tests decide what ships. Your $20 plan stops hitti

AiSOC, güvenlik telemetrisini birleştirip tehditleri tespit eden ve AI agent ile incelerken karar adımlarını kayda alan, self-hosted bir SOC platformudur.

Neden yükseliyor · 149 günlük yaşına rağmen 2.370 yıldız ve son 28 günde 247 commit dikkat çekiyor; self-hosted kurulum, yerel modelle API key gerektirmeyen başlangıç ve çalışır demo akışı projeyi denemeyi kolaylaştırıyor. Commit sayısı yüksek olsa da katkı 3 farklı yazarla sınırlı; yükselişin nedenini doğrudan doğrulayacak benimsenme verisi yok.

Claude Code ve Codex ile AI destekli geliştirmeyi yapılandırmak için promptlar, şablonlar ve kalite kontrol pratikleri sunan bir toolkit.

Neden yükseliyor · 44 günlük bir repoda 729 yıldız ve son 28 günde 14 commit dikkat çekiyor. AI coding agent'lar için uygulanabilir iş akışı ve kalite kapıları sunması ilgi çekebilir; ancak commit'lerin tek yazardan gelmesi bağımsız katkı ve doğrulama kanıtını sınırlıyor.

Cost-efficient coding-agent runtime: prefix-aware prompt-cache routing, poisoning-safe semantic cache, budgeted context compression, request

SkillCorpus, dağınık SKILL.md dosyalarını agent'lar için seçilmiş, aranabilir skill corpus'larına dönüştüren ve retrieval ile evaluation araçları sunan açık kaynak bir altyapıdır.

Neden yükseliyor · Agent'lar için paylaşılan ve göreve göre seçilen skill katmanına ilgi var; repo kısa sürede 667 yıldıza ulaşmış ve son 28 günde 9 commit ile 3 farklı yazar katkı vermiş. README ayrıca birden fazla benchmark'ta iyileşme bildiriyor; bu sonuçları bağımsız olarak doğrulayamıyorum.

[Public preview] Externally scored agentic ML research benchmark: 60 tasks, real competition ground truth. Open protocol, operated evaluatio

Official Code for the paper "Doc2Harness: Benchmarking LLMs on Long-Horizon Harness Engineering of Coding Agents "

The first exam for code agents in the physical world: frozen robot tasks from many benchmark suites, each answered with a program and verifi

[EMNLP 2026] MobileGym: A Verifiable and Highly Parallel Simulation Platform for Mobile GUI Agent Research · 浏览器里运行的安卓模拟器 · Browser-hosted A

OrcaReplay, AI agent çalıştırmalarını kaydedip çevrimdışı tekrar oynatmaya ve bir adımdan farklı bir modelle devam ettirmeye yarayan bir debugging aracıdır.

Neden yükseliyor · 29 günlük bir repo için 171 commit ve 5 farklı yazar kayda değer geliştirme etkinliği gösteriyor; 268 yıldız da erken dönem ilgisine işaret ediyor. Yükselişin belirli bir duyuruya veya kullanıcı benimsemesine bağlı olup olmadığı verilen verilerden anlaşılamıyor.

Co-manage memory with your coding agent across sessions. A local workbench for Claude Code, Codex and Pi.

Your agent’s trace says “success” but the answer is wrong. Tracedown finds the step that caused it and replays the run with a fix.

Code-only runtime toolkit for cost-aware multi-agent organization and control

Local-first agent platform: board + role agents + agent CLI runs (Claude Code, Cursor, Antigravity, OpenCode) or local and API models (Ollam

The why layer of repo-native project memory: the reasoning behind a codebase as Markdown in the repo, versioned by Git, for coding agents an

Self-hosted AI workspace with agents, skills, and tools (Gmail, Calendar) that runs entirely on your own provider API keys (BYOK). Bring you

Can coding agents build the agent you actually want? A benchmark of 25 agent software engineering tasks across creation, editing and optimiz

Universal, model-agnostic operating harness for AI agents (Claude, Codex, Gemini, …) — a lean core + work-type profiles assembled by one set

Manipulation benchmarks test acting to change; EmbodiedLab tests acting to know. Given only an objective and the apparatus, a robot runs its

A recursive self-improvement (RSI) workflow that evolves both target agents (e.g., Pi and DeepSeek Harness) and the RSI system itself.

Python toolkit for PowerPoint ↔ HTML, built for AI. Round-trip fidelity preserves formatting across every edit loop — ideal for RAG ingestio

Give AI agents ambitious work without losing the plot. nac is an open-source harness for long-running tasks, using a central orchestrator, t

An agent skill that plays ARC-AGI-3. One rule: say what an action will do before you spend it. Claude Code on Opus 5 finished all 25 public

DeepSeek Harness 的执行迷宫——看 Agent 真实怎么干活：迷宫时间轴 · 数据轨道 · 确定性执行分析 · 多会话对比 | The execution maze for DSH agents: maze timeline, per-step data trac

A linked-list discipline for knowledge, memory and pointer hygiene in large projects: five laws plus portable Python engines (declarative ru

Multi-Agent AI Task Orchestrator 2026

A custom coding harness for TypeSafe AI's Jev: an LLM proposes, Jev answers narrow questions, code decides, every step leaves a receipt.

Turn your repos into a team of Claude Code agents with persistent memory, coordinated tasks, and a live dashboard.

Offline-first Python AI agent that runs a tiny research business: quotes each job against its own costs, collects via Stripe, fulfils with N

DrugTargetWorld: a synthetic biobank for training and benchmarking AI scientists on end-to-end drug target discovery. Paper: arXiv:2610.0955

A task conductor for Claude Code that gets better the more you use it. Multi-agent orchestration with an evidence contract that BLOCKS inste

Project memory for long-running AI work across agents, models, and sessions. Keep context, decisions, progress, and next steps in local proj

An unattended maintainer for one repository. It reviews, files, prunes, and merges only against proof that a guard bites.

22 production-grade skills for AI coding agents — setup, scope, discipline, code review, security, session management, governance, ops, and

SciPIvot: Benchmarking Principal-Investigator Judgment at Pivotal Moments of Scientific AutoResearch Agents

Swap, lend, trade perps, add liquidity, move SOL, read the market. One binary your agent calls, across every venue.

SimuVerity: Benchmarking Agents for Engineering-Grade Simulink Model Generation

Kit is a coding agent runtime. Claude Code but faster, cheaper and more to the point 🎯

十三希诺 Agent 工作流 · 中文源库：把工程纪律做成可机检门禁的元工作流 Skill 家族——三级跑道／判级／GATE 12 字段可复跑证据／405 条细则库／五包交付（核心+流程+角色+产品工程+单文件蒸馏），五平台注入适配。docs/ 为开发工作区，实验数据已归档。En

Reproduction package for "Harness as a Language: A Minimalist Agent Framework With Maximal Expressivity" — ten agent arms across AppWorld an

Local outcome-loop + reward layer for coding agents — remember what worked, ship the next move, measure how it landed. Docs at nimroboai.com

The Open-Source AI Career Operating System. 167 Specialized AI Agents across 19 Divisions, Spoken Voice Lab (27 Languages), AI job search ,A

deedchain, browser agent raporlarındaki iddiaları kaydedilmiş ground truth ile karşılaştırarak rapor doğruluğunu ölçen bir benchmark ve Python kütüphanesidir.

Neden yükseliyor · Repo yalnızca 23 günlük ve 28 yıldızlı; son 28 gün commit ve farklı yazar verisi ölçülmediği için yükselişin arkasındaki aktivite veya benimsenme düzeyi doğrulanamıyor. Rapor doğruluğunu ayrı bir değerlendirme konusu yapması dikkat çekici olsa da README'deki “first benchmark” iddiası bağımsız olarak doğrulanmış değil.

The agent that evolves with you 𖤍

A sandbox service for running untrusted, agent-authored code that reports the isolation boundary each run executed behind. Lessons, no clone

SWE-Serve: an agentic benchmark of 53 production inference-engineering tasks derived from merged SGLang pull requests, run with Harbor.

Skill to generate the knowledge vault for projects using the Ralph loop

Let Claude Code, Codex and Pi talk to each other: cross-session messaging that wakes AI agents on one machine or across your LAN (macOS/Linu

The control plane for your coding agents, however you run them. One checkout of rules, skills, subagents, hooks and stances, projected into

SDK, skills, and examples for Inkbox: give AI agents a persistent identity with email, phone number, and internet address