AI Agents

DeepSeek Harness v0.1 is an MIT-licensed agent harness where every capability is a Cordis plugin. Four runtime modes, append-only session logs, and provider-agnostic model routing.

Meta's Muse Glimmer is a 30B open-weights agentic model under Apache 2.0. It fits 24 GB VRAM and decodes 3.1x faster with DFlash speculation.

Long agent runs accumulate state that no transcript records — edited files, a live dev server, installed packages, a warm prompt cache. When an agent misreads a traceback at step 10 and rewrites a correct file, patching forward burns tokens and restarting re-pays every call. Researchers at Northeastern University and Stanford University released Shepherd, an MIT-licensed Python runtime substrate that records every agent-environment interaction as a typed event in a Git-like execution trace. Each commit covers the agent process and filesystem together, copy-on-write, so a rewind restores live state instead of just files. The paper reports 5× faster forks than Docker, over 95% prompt-cache reuse on replay, and a live supervisor raising CooperBench pair-coding pass rates from 28.8% to 54.7%.

Tencent Cloud has open-sourced TencentDB Agent Memory v2.0, a team-level memory hub that turns conversations, documents and code into four governed, reusable assets — Chat Memory, Skill, LLM-Wiki and Code-Graph. It is MIT-licensed, self-hosted via Docker, and integrates with Claude Code, OpenClaw, Hermes and CodeBuddy. The differentiator is not retrieval but governance: ACL-based visibility decides which agent gets which asset, and which version is valid.

Microsoft has open sourced code-testing-generator, a polyglot unit-test agent shipping in the MIT-licensed dotnet/skills repository. It reads a repository before writing anything — detecting the language, test framework, existing conventions, and the real build and test commands — then plans, writes, runs and validates the tests it produces. On Microsoft's internal 152-task benchmark it completed 140 tasks against 120 for stock GitHub Copilot on the same model, with the gain concentrated almost entirely in vague prompts and diff-targeted requests.

Cloudflare has released Kitesurf, a stateless web browser built specifically for AI agents that runs entirely in V8 isolates on Cloudflare Workers, with no Chromium underneath. The browser drops human-facing features like tabs and extensions in favor of what agents need: machine-readable content, scalability, and isolation. Built in 12 weeks using Rust components like Blitz, Stylo, and Boa JS, it already passes 215,000+ Web Platform Tests. Benchmarks show 3.1–3.8× less CPU and 4.7–7.0× less memory than Chromium on screenshots and HTML extraction. Existing Puppeteer, Playwright, and MCP clients work by adding a single browser=kitesurf parameter, free while in beta.

Prime Intellect has open-sourced Prime Agent, a coding and research harness built on two abstractions: the Recursive Language Model, which turns sub-agent calls into functions inside a persistent IPython kernel, and the Continual Harness, which lets the agent edit its own prompts, skills, memory, and sub-agent specs mid-run. With Opus 5 it reports 95.5% RHAE Best@1 on ARC-AGI-3, above the reported human expert baseline of 95.4%.

Most coverage of Microsoft's SkillOpt centers on its 52/52 result. The more consequential finding is in Section 4.3: the exported best_skill.md keeps working in environments it was never trained on. A Codex-trained SpreadsheetBench skill lifted Claude Code from 22.1 to 81.8, slightly above the 80.4 that harness reached training its own skill. Retention varies sharply by task type — 102% on spreadsheets, 10% on math — which is what makes the result worth reading closely.

Meta Superintelligence Labs has released Muse Code, a terminal coding agent in beta, powered by the new Muse Spark 1.2 model. Muse Code plans changes, writes code, and validates results across large repositories. Async background agents stay active for the whole session instead of spawning per task. A local append-only event log makes the runtime replay-exact and restart-safe after a crash. Muse Spark 1.2 was co-trained with the harness and trained on long-horizon, repository-scale work.

CopilotKit has published the Channels SDK, an MIT licensed library that runs an existing AG-UI agent inside Slack and Microsoft Teams. Version 0.5.0 ships five platform adapters and a documented runtime contract. This breakdown covers the verified deployment paths, the baseline requirements, and the one dependency that is easy to miss

Learn how to build an end-to-end security assessment pipeline for AI agent skills using NVIDIA SkillSpector and LangGraph. In this tutorial, we construct a synthetic skill marketplace, scan for malicious prompt injection, credential access, and risky dependencies, and implement custom YARA rules, baseline suppressions, and CI deployment gates.

Y Combinator has open-sourced QM, the multiplayer agent harness it uses internally across accounting, legal, events, and engineering. Released July 31, 2026 under an MIT license, QM gives each employee an isolated workspace and each Slack room its own scoped memory, files, keychain view, permissions, crons, web apps, and durable sandbox. Pi, OpenCode, Codex, and Claude Code all drive the same headless core, so deployments avoid vendor lock-in.

Supabase has open sourced supabase/evals, an Apache-2.0 benchmark and framework that runs coding agents including Claude Code, Codex and OpenCode against real Supabase tasks — building schemas, debugging Edge Functions, fixing RLS policies — inside containerized stacks, then scores them with deterministic checks and LLM-as-a-judge.

Three terms now compete for the same line in AI engineering job descriptions. Prompt engineering is the established one. Loop engineering entered the AI...

In this tutorial, we build an advanced workflow around Anthropic’s financial-services repository and reproduce its skill-driven architecture in pure Python. We begin by installing...

Perplexity has released pplx, an official command line client for its Search API. The tool exposes two commands — pplx search web and pplx content fetch — and returns exactly one JSON object on stdout. It ships as a checksum-verified single binary for macOS arm64 and Linux, alongside an Agent Skill for Claude Code, Codex CLI and any harness that can read a URL.

OpenAI disclosed that its own models breached Hugging Face's production infrastructure while taking a public security benchmark. The models were not attacking a target — they were optimizing a score. Here is the mechanism, what the ExploitGym data showed two months earlier, and which widely repeated claims about the incident are not actually confirmed.

Discover how to create self-evolving AI agents using the OpenSpace framework. This tutorial guides you through the entire workflow—from environment setup and custom skill creation to MCP integration and using SQLite to manage agent lineage—empowering you to build more efficient, reusable agent systems.

Andrew Ng has released OpenWorker, an MIT-licensed desktop AI agent that returns finished deliverables instead of chat replies. It runs a local Python agent server under a Tauri shell, supports 30 curated tool-calling models plus fully local Ollama, and gates every write, shell command and off-machine action behind a typed risk engine.

Meta has open-sourced Astryx, the React and StyleX design system it ran internally for eight years across 13,000+ apps. It ships 150+ accessible components, seven themes, dark mode, templates, and an agent-ready CLI under MIT — with React 19+ required.

Google Cloud's generative-ai repository ships the Always-On Memory Agent, a reference implementation that treats memory as a running process. Built on Google ADK and Gemini 3.1 Flash-Lite, it uses no vector database and no embeddings. Instead, an orchestrator routes to Ingest, Consolidate, and Query sub-agents that read, connect, and write structured memory into SQLite 24/7.

Introduction This tutorial starts where most agent demos stop: giving the agent persistent memory, operational context, and a place to write back what happened. An...

We explore the Patter SDK by building a voice-agent workflow for a restaurant booking use case. We define dynamic caller variables, register callable tools for availability, bookings, hours, and human transfer, and layer output guardrails over every reply. We simulate speech-to-text and text-to-speech behavior, run scripted call flows, and track modeled latency and cost in a dashboard. We validate the agent with a deterministic eval harness, then map the same logic to a real deployment using Twilio and OpenAI Realtime.

SpaceXAI published the Grok Build source on July 15, 2026. The Apache 2.0 Rust tree covers the agent loop, tool dispatch, the TUI, and the extension system. Grok 4.5 stays closed, and external contributions are not accepted.

See how Vibe, Claude Code, Cursor, and Codex compare on cost, open weights, self-hosting, and async agent surfaces.

In this tutorial, we reconstruct the VideoAgent workflow as a runnable, API-key-free multi-agent pipeline. We build an intent parser, an agent library, a tool router, a graph planner, and a textual-gradient optimizer that repairs the execution graph. We wire these planning components to FFmpeg, Whisper transcription, scene detection, keyframe sampling, captioning, cross-modal indexing, and beat-synced editing. By the end, we have a system that answers questions about a video, summarizes it, and produces edited artifacts from natural-language instructions.

OpenAI moved GPT-5.6 to general availability on July 9, 2026, shipping three tiers instead of one model. Sol is $5/$30 per 1M tokens, Terra is $2.50/$15, and Luna is $1/$6. Sol sets the Artificial Analysis Coding Agent Index at 80, 2.8 points above Claude Fable 5, and reaches 62.6% on OSWorld 2.0 using 85% fewer output tokens than Opus 4.8. The substantive developer change is Programmatic Tool Calling, which runs model-written JavaScript in an isolated V8 runtime to orchestrate tools without returning every intermediate result to the model. Clio reports a 38% cut in prompt tokens, PlayCo 63.5% fewer total tokens. The gaps are real too: Fable 5 still leads the Artificial Analysis Intelligence Index, GDPval-AA v2, and Toolathlon, and Claude Mythos 5 leads SWE-Bench Pro by roughly 15 points.

SpaceXAI released Grok 4.5, a Cursor-trained model for coding, agentic tasks, and knowledge work. It serves at 80 TPS, costs $2/$6 per million tokens, and ranks #1 on Harvey's Legal Agent Benchmark.

Junyang Lin, the former technical lead of Alibaba's Qwen, walked through the model family in a talk "towards a generalist model / agent," then expanded it in an essay. We read both for practitioners: Qwen3 hybrid thinking modes and dynamic thinking budgets, where the merge fell short, the shift from reasoning thinking to agentic thinking, why agentic RL infrastructure is harder, and where reward hacking bites.

A hands-free NVIDIA agent framework hosts each RTL problem as a versioned repository, reaching 100% completion across benchmarks.

Mistral AI released Leanstral 1.5, a free Apache-2.0 code agent model for Lean 4. It saturates miniF2F and solves 587 of 672 PutnamBench problems. The 119B mixture-of-experts activates 6.5B parameters per token. We break down its architecture, benchmarks, real bug-finding case studies, and deployment code.

WebBrain is a free, MIT-licensed AI browser agent for Chrome and Firefox. It reads pages, extracts data, and automates multi-step tasks through Ask and Act modes. Run it on local models like llama.cpp or Ollama for privacy, or connect any cloud API.

Alibaba's Page Agent runs as client-side JavaScript inside the webpage. It reads the live DOM as text, then clicks and types from natural-language commands. No screenshots, no multimodal model, and no backend rewrite are required.

Linq launches iMessage Apps: interactive imessage_app cards that run payments, tickets, flights, and games inside the iMessage thread for agents.

OpenClaw's iOS and Android apps are companion nodes, not standalone chatbots. Each phone pairs to a self-hosted Gateway over WebSocket. This adds device hardware — camera, location, voice, and Canvas — to a local-first AI agent. Here is the architecture, the capabilities, and the trade-offs for builders.

NVIDIA's open-source BioNeMo Agent Toolkit turns biomolecular models like OpenFold3, DiffDock, and GenMol into documented, callable skills for AI agents. Each skill describes a model's purpose, inputs, artifacts, and failure modes, so an agent can select, run, and interpret it. In NVIDIA's benchmarks with Codex CLI and GPT-5.5 fast, skills raised task completion from 57.1% to 100% and doubled token efficiency.

EverMind has open-sourced EverOS, a local-first memory runtime that stores AI agent memory as plain Markdown indexed by SQLite and LanceDB. It combines hybrid BM25 + vector retrieval, multimodal ingestion, and self-evolving Skills under an Apache 2.0 license. Here's what it is, how the architecture works, where the benchmarks stand, and where it still falls short — plus a runnable code walkthrough and an interactive demo.

Meta released Astryx, an open-source React design system built on StyleX. It pairs a CSS-variable theme cascade with a CLI and MCP server, so both engineers and AI agents build using the same API. The project is in Beta, MIT-licensed, and grew inside Meta over eight years.

A Cursor study shows coding agents retrieve known fixes instead of deriving them, inflating SWE-bench Pro scores through runtime contamination.

Nous Research has added /learn to the Hermes Agent Skills System. The command authors a standards-compliant SKILL.md from a local directory, a doc URL, a past conversation, or pasted notes. The live agent sources the material with its own tools, then writes the skill — no hand-writing and no separate ingestion engine. Here is how it works, where it fits, and what to review before trusting the output.

Fugu and Fugu Ultra route tasks across a swappable model pool, leading most coding, reasoning, and agentic benchmarks.

LLMs are stateless by default. Agent memory fixes that. This guide breaks down all 7 types — working, semantic, episodic, procedural, retrieval, parametric, and prospective. It covers what each stores, where it lives, and when to build it. Includes a comparison table and working Python code.

Nous Research has added a Blank Slate setup mode to its open-source Hermes Agent. It starts an agent with everything off except provider, model, File Operations, and Terminal. You opt in to the rest.

Perplexity has launched Brain, a self-improving memory system for its Computer agent. Instead of remembering the user, Brain remembers the agent's work — what worked, what failed, and what corrections got made. It builds a traceable context graph, reviews it overnight, and reports early gains in correctness, recall, and cost.

In this tutorial, we use NVIDIA SkillSpector to evaluate AI skills for security risks before deployment. We build a corpus of benign and deliberately vulnerable skills, then scan them through SkillSpector's programmatic LangGraph workflow. We organize the risk scores and findings with pandas, then visualize severity and category distributions. We export results in SARIF format, register a custom analyzer, and optionally apply an LLM-based semantic pass.

We look at Hermes Agent's new asynchronous subagents from Nous Research. The delegate tool can now spawn background agents that no longer block the parent chat. We walk through the async_delegation toolset tracked in issue #5586, covering how to spawn, check, steer, and collect tasks. We also cover subagent isolation, the synchronous-versus-asynchronous tradeoffs, and when to reach for ACP instead

Claude Code is a layered agentic coding tool, not a single chat prompt. This guide breaks down 25 features, from CLAUDE.md, skills, subagents, and hooks to MCP and Auto Mode. It includes a comparison table, working code examples, real use cases, and an interactive demo you can try.

Databricks has open-sourced Omnigent, a meta-harness that sits above coding agents like Claude Code, Codex, and Pi. It adds composition, contextual policies, and live session sharing under one interface, on terminal, web, desktop, and mobile. The Apache 2.0 project is in alpha.

In this tutorial, we implement a QwenPaw workflow that provides a practical environment for building and testing an agent-powered assistant. We install and initialize...

shutdown followed a US government export control directive citing national security authorities. All other Anthropic models, including Opus 4.8, remain available.

Moonshot AI has open-sourced Kimi K2.7-Code under a Modified MIT license. It is a coding-focused, agentic model built on Kimi K2.6, with a 256K context window and roughly 30% lower reasoning-token usage. Moonshot reports gains over K2.6 on six benchmarks, including +21.8% on Kimi Code Bench v2. The model is available via the Kimi API and Kimi Code.

Moonshot AI's Kimi Work is a local desktop agent for macOS and Windows. It runs a 300-sub-agent swarm, drives your logged-in browser via WebBridge, and schedules background jobs.

Grok Build's in-terminal marketplace bundles skills, agents, hooks, and MCP servers, with commit-SHA verification on every remote plugin.

The Hermes Agent dashboard now builds complete agent profiles in one flow, replacing multi-step CLI setup for users.

Software development has changed. Engineers no longer type most code by hand. They describe intent, and AI agents do the work. Modern tools plan...

UIUC and Chroma's Harness-1 is a 20B retrieval subagent trained with reinforcement learning inside a stateful search harness. The harness maintains the bookkeeping — candidate pool, importance-tagged curated set, evidence graph, verification records — while the policy decides what to search, curate, verify, and when to stop. It reaches 0.730 average curated recall across eight benchmarks, beating the next open subagent by 11.4 points and trailing only Opus-4.6. Weights and harness code are public.

Google released the Colab CLI, letting developers and AI agents run local code on remote Colab GPU and TPU runtime

Kimi Code CLI is Moonshot AI's open-source terminal coding agent, written in TypeScript with subagents and MCP configuration.

Vibe coding turns plain language into working software. Explore 15 tools shaping how developers build apps in 2026.

Hermes Desktop is a no-terminal GUI sharing one agent core, skills, and memory with the Hermes Agent CLI.

Describe a dataset in one sentence; Bigset's orchestrator and parallel sub-agents research the live web and return structured tables.

MiniMax M3 introduces MiniMax Sparse Attention, a 1M-token context window, and native image, video, and computer use support.

Nous Research's Hermes Agent adds Tool Search to fix MCP context bloat using BM25 progressive schema disclosure.

Hexo Labs released SIA, an open-source self-improving loop, under an MIT license. A Feedback-Agent reads each run's trajectory, then either rewrites the scaffold or triggers a LoRA weight update on gpt-oss-120b. Combining both levers beat scaffold-only iteration on LawBench, TriMul GPU kernels, and scRNA-seq denoising.

Anthropic's Claude Opus 4.8 brings dynamic workflows and cheaper fast mode to Claude Code, now in research preview

Microsoft Research introduces Webwright, a terminal-native browser agent framework that replaces click-trace web automation with reusable Playwright scripts. Using a single agent loop across three modules and roughly 1,000 lines of code, Webwright powered by GPT-5.4 reaches 60.1% on the long-horizon Odysseys benchmark and 86.7% on Online-Mind2Web — the highest AutoEval score among open-sourced harness recipes.

In this tutorial, we build an advanced workflow using the SuperClaude Framework as a structured layer on top of the Anthropic API.

An inside look at CopilotKit’s 2026 shipping cycle. Learn how the new AG-UI protocol, AIMock testing suite, and Pathfinder server are providing the production architecture developers need for agentic AI.

Not all database platforms are built for the same job.Not all database platforms are built for the same job. Here is how Upstash, Supabase, and Neon actually differ — and which one fits your vibe coding workflow in 2026.

Enterprise agentic AI has moved from pilots to production in 2026. This guide ranks the top 10 platforms — Salesforce Agentforce, Microsoft Copilot Studio, ServiceNow, LangGraph, and more — with verified pricing, real adoption data, and honest constraints to help enterprise teams make the right platform decision.

Vercel Labs has released Zero, an experimental systems programming language designed so AI agents can read, repair, and ship native programs without requiring human interpretation of compiler output. The language emits JSON diagnostics with stable codes and typed repair metadata, enforces capability-based I/O at compile time, and compiles to sub-10 KiB native binaries.

The AI coding agent field in 2026 is more capable, more fragmented, and harder to benchmark than it looks. Claude Code leads on code quality at 87.6% SWE-bench Verified. GPT-5.5 tops Terminal-Bench at 82.7%. But the benchmark OpenAI itself declared contaminated in February 2026 is still being used to rank these tools — including by the labs publishing their own scores.

Cline has extracted its internal agent harness into an open-source TypeScript SDK called @cline/sdk, the same runtime now powering its CLI and Kanban, with VS Code and JetBrains extensions being migrated. The SDK is structured as a four-layer stack — @cline/shared, @cline/llms, @cline/agents, and @cline/core — with native support for plugins, subagents, CRON scheduling, checkpointing, and MCP connectors. On Terminal Benchmark 2.0, Cline CLI scored 74.2% on claude-opus-4.7, compared to Anthropic's published 69.4% for Claude Code on the same model. Install via npm install @cline/sdk. Requires Node.js 22+.

Thinking Machines Lab has introduced a research preview of TML-Interaction-Small, a 276B parameter Mixture-of-Experts model with 12B active parameters, built around a multi-stream, time-aligned micro-turn architecture that processes 200ms chunks of audio, video, and text simultaneously — eliminating the need for external voice-activity detection harnesses. Unlike standard turn-based models that freeze perception during generation, the system runs two components in parallel: a real-time interaction model that maintains continuous full-duplex exchange with the user, and an asynchronous background model that handles sustained reasoning and tool use while sharing the full conversation context throughout.

Hermes Agent, the open-source self-improving AI agent from Nous Research, has overtaken OpenClaw to claim the #1 position on OpenRouter's global daily token rankings as of May 10, 2026 — generating 224 billion daily tokens versus OpenClaw's 186 billion. The milestone places a Nous Research project ahead of an OpenAI-sponsored platform in real-world daily inference volume, just three months after launch.

If you have spent time using AI coding agents — GitHub Copilot, Claude Code, Gemini CLI — you have probably run into this situation:...

OpenAI has shipped a Chrome extension for Codex, its AI coding agent, enabling it to complete browser-based tasks directly inside Google Chrome on macOS and Windows — including interacting with signed-in websites, using Chrome DevTools, and running multi-step workflows across browser tabs.

In this tutorial, we explore CloakBrowser, a Python-friendly browser automation tool that uses Playwright-style APIs within a stealth Chromium environment. We begin by setting...

In this tutorial, we build a Groq-powered agentic research workflow that runs directly using Groq’s free OpenAI-compatible inference endpoint

CopilotKit Intelligence adds a managed persistence layer on top of the open-source CopilotKit stack, giving agents the ability to retain context, state, and interaction history without custom storage infrastructure

Discover the top search and fetch APIs for AI agents in 2026. Compare tools like TinyFish, Tavily, and Firecrawl based on latency, token efficiency, and free tiers to optimize your agent's web retrieval.

Build a Multi-Agent AI Workflow for Biological Network Modeling, Protein Interactions, Metabolism, and Cell Signaling Simulation

Cursor Launches TypeScript SDK to Let Developers Build and Deploy Programmatic Coding Agents

In this tutorial, we build a Reinforcement Learning–driven agent that learns how to retrieve relevant memories from a long-term memory bank. We start by...

As AI agents move from research demos to production deployments, one question has become impossible to ignore: how do you actually know if an...

A new memory framework from Google Cloud AI Research and UIUC gives LLM agents the ability to distill generalizable reasoning strategies from both successful and failed experiences — and combines that with test-time scaling to create agents that genuinely improve over time.

How to make multiple agents work together like an elite team — autonomously dividing tasks, communicating efficiently, and collaborating seamlessly? The openJiuwen community released the...

Debugging an AI agent that runs for dozens of steps: reading files, calling APIs, writing code, and revising its own output, is not like...

Moonshot AI, the Chinese AI lab behind the Kimi assistant, today open-sourced Kimi K2.6 — a native multimodal agentic model that pushes the boundaries...

In this tutorial, we build a universal long-term memory layer for AI agents using Mem0, OpenAI models, and ChromaDB. We design a system that...

In this tutorial, we build an advanced, production-ready agentic system using SmolAgents and demonstrate how modern, lightweight AI agents can reason, execute code, dynamically...

Google just announced the release of Skills in Chrome, a new feature built into Gemini in Chrome that lets users save frequently used AI...

AI agents struggle with tasks that require interacting with the live web — fetching a competitor's pricing page, extracting structured data from a JavaScript-heavy...

In this tutorial, we build an advanced data analysis pipeline using Google ADK and organize it as a practical multi-agent system for real analytical...

Standardized tests can tell you whether a student knows calculus or can parse a passage of text. What they cannot reliably tell you is...

MiniMax, the AI research company behind the MiniMax omni-modal model stack, has released MMX-CLI — Node.js-based command-line interface that exposes the MiniMax AI platform's...

MiniMax has officially open-sourced MiniMax M2.7, making the model weights publicly available on Hugging Face. Originally announced on March 18, 2026, MiniMax M2.7 is...

In this tutorial, we build and operate a fully local, schema-valid OpenClaw runtime. We configure the OpenClaw gateway with strict loopback binding, set up...

Training AI agents that can actually use a computer — opening apps, clicking buttons, browsing the web, writing code — is one of the...

There's a particular kind of tedium that every AI engineer knows intimately: the prompt-tuning loop. You write a system prompt, run your agent against...