Havuz

CatchBench: a benchmark for auditing agent failures across the full lifecycle.

20 runnable LLM agent design patterns in Python, benchmarked, traced, offline-compatible. ReAct, multi-agent, constitutional AI, and more.

AI World Generator 2026: Create Self-Evolving Maps & Stories

Multi-Agent AI Task Orchestrator 2026

Single-agent multi-modal system for insurance damage claim verification. #1 globally, HackerRank Orchestrate June 2026 (15,295 registrants).

Don't trust an autoresearch paper at face value. Reviewer-side integrity forensics (self-consistency + fabrication), deterministic verdict.

🔥 CogitoAgent — An autonomous AI agent, cloud-driven but locally executed. It continuously thinks, automatically analyzes, and integrates pr

Context optimization for AI coding agents — lower token cost and latency without sacrificing accuracy.

The agent that learns how you decide - and proves it by never deciding for you. A local-first, loop-native runtime for a team of AI agents o

Agentic Runtime

Durable, searchable memory of your past agent sessions.

Claude Code plugin for autonomous AI research — multi-agent loops take a bare topic all the way to running experiments, with no human-writte

A constitution generator for autonomous, agent-run projects: an interview-driven meta system prompt (and installable skill) that scaffolds a

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

An open-source research system for autonomous investigation, recursive self-improvement, and human–AI collaboration. Run experiments, refine

Where agents build agents. An open-source Rust framework for complex agentic systems: ontology-grounded, auditable, one binary.

Audit any agent decision across its past, present, and future, on one typed graph.

Decision provenance and data-flow tracking for coding agents: computes what changed in your code and data, records why it changed, and surfa

An autonomous agentic pipeline that finds, proves, and patches real C memory-safety vulnerabilities end-to-end using a single 7B open-source

Loop until it's better — drop-in agentic loops (autoresearch, scientific writing, data analysis, code/SQL/prompt optimization, red-teaming)

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

Long-term memory for AI agents with no LLM call: an MCP server and Claude Code plugin, plus local-first context compression for LLM apps.

EMNLP 2026 Industry Track | A benchmark for shopping agents on long-horizon tasks with distributed hidden intent.

Runtime security for AI agents. In-process, zero dependencies, Apache 2.0.

Readable agent skills with optional declarative YON protocols you can inspect and validate. A personal, field-used pack by Alexandru Mares.

Reusable multi-round research skill for evidence-backed product, market, technical, open-source, and investment-style surveys.

Train computer-use agents end-to-end on real desktops, at scale: a high-performance runtime where 8K+ live environments on one laptop thread

Deterministic code-graph (GraphRAG) over your repo for LLM agents — local-first, git-native, zero-infra, served via MCP. Python, TS/JS, Rust

一个将现实议题、规则机制与角色关系转化为可交互多智能体仿真/演绎世界的生成与运行框架。

Open-source alternative to Claude Co-Work built by Claude. with BYOK

EurekAgent: an autonomous research system for metric-driven tasks, built with Claude Code. Define the problem and metric. Get breakthrough r

Claude Code plugin that makes coding agents measurably cheaper over time: collect token costs, distill candidate rules, benchmark them on a

Catch your AI agents when they lie about what they shipped — verifies claims against git instead of believing the agent.

A next-generation automatic algorithm design platform, making automated algorithm design more accessible and easier to use

Multi-agent demo for Statewave conflict resolution. Concurrent agents write to a shared subject; the compiler detects contradicting facts vi

Chromium build for AI agents and scrapers: 33 agent-level CDP commands, stable fingerprint identities, free Windows binary.

RHO: Evolving Agents in the Dark — Retrospective Harness Optimization via Self-Preference. Improving LLM agents from unlabeled past trajecto

EC2-Bench (EC²-Bench): benchmark for edge-cloud collaboration in long-horizon LLM agents — task performance, resource usage & privacy exposu

A local CLI for logging predictions before the outcome and checking whether your "80% sure" comes true 80% of the time.

Can LLM/MARL agents coordinate in long-horizon, open-ended tasks? NeurIPS 2026 (Evaluations and Datasets).

Forkable shell substrate for a single code repository — DB-backed identity, memory, roadmap, content; harness-agnostic boot. We build the da

Models Take Notes at Prefill: KV Cache Can Be Editable and Composable (arXiv:2606.17107) — paper, code, results, and interactive companion.

Multi-model consensus debate via the filesystem. LLMs propose, peer-review, rebut, vote and synthesize a group-confirmed answer. CLI + MCP.

A living AI-agent office on your desktop wallpaper — Claude Code agents that walk, work, delegate, learn & hold meetings. Per-agent swappabl

Open-World Self-Evolution for LLM Agents — agents that build both their skills and their own verification signals from scratch, with no targ

Intent-bound action authorization for AI agents — deterministic, per-action approval a prompt injection can't re-point, with a signed, tampe

Counterfactual replay and causal attribution over LLM-agent trajectories: find which step actually caused a failure by intervening on it.

PaperJury, Claude Code içinde araştırma makalelerini LaTeX üzerinde gözden geçirip sınırlı ve kontrollü düzeltmeler önermek için kullanılan bir skill'dir.

Neden yükseliyor · 117 günde 1207 yıldız alması; arXiv makalesi, dogfood örneği ve sayısal baseline karşılaştırmalarıyla dikkat çekiyor. Ancak son 28 günde commit ve farklı yazar olmaması, yükselişin güncel geliştirme etkinliğiyle desteklendiğine dair kanıt sunmuyor.

Forge your own agent harness — a config-to-code generator that renders a thin, standalone Python agent harness you fully own. No LangChain/L

Helix is ​​a powerful self-learning AI agent with memory, skills, and the ability to recall tools. It learns from successful tasks and creat