📂 AI Agents
How to create, configure and deploy autonomous AI agents to automate your repetitive tasks and increase your productivity.
GitHub unveils the Agent Host Protocol: code agent sessions that survive device switches
Discover GitHub's Agent Host Protocol: your coding agent sessions survive closing the editor and follow you across all your devices.
Google /teamwork proves the Courtade-Kumar conjecture: agent swarms enter formal mathematics
Google /teamwork proves the Courtade-Kumar conjecture, an open problem since 2013: Gemini 4 agent swarms enter formal mathematics.
ChatGPT Wallet: OpenAI launches agentic payment with Visa, Mastercard, and Amex — the 'dots' agents buy for you
ChatGPT Wallet: OpenAI launches agentic payments with Visa, Mastercard and Amex. The "dots" agents buy for you, from discovery to payment.
Meta launches Muse Ember: a multimodal model at $0.10/M tokens for running agent swarms
Meta launches Muse Ember, a $0.10/M-token multimodal model optimized for agent swarms: tool calls, code, routing, and data extraction.
claude-mem: 114,000 stars in just a few weeks for persistent agent memory
claude-mem: the open source plugin that gives coding agents persistent memory. Discover why it's surpassing 114,000 stars on GitHub.
ServerBench: when your coding agents design the very infrastructure that runs them
ServerBench: coding agents design the very LLM infrastructure that runs them, upending the roles of the vLLM, SGLang and TensorRT-LLM runtimes.
EvoTeam: evolving the team of agents rather than the agent itself
EvoTeam: an approach that evolves the team of AI agents rather than the agent itself, to solve failures of multi-agent systems.
Anthropic: 86% of analysts follow their agent's hidden slope, and 77% repeat its arguments without ever citing it
Anthropic study: 86% of analysts unknowingly follow their AI agent's opinion. What this changes for enterprise agent deployment.
Anthropic hands Agent Skills to the Linux Foundation: the skills standard moves under neutral governance, just as MCP did before it
Anthropic hands Agent Skills to the Linux Foundation: the AI skills standard moves under neutral governance via the AAIF, like MCP before it. Analysis &...
Cursor Computer Use: the agent steps out of the browser and drives any desktop app, including the iOS Simulator
Cursor Computer Use: the coding agent leaves the browser and drives your desktop apps, from the iOS Simulator to your daily tools. Discover this revolutio
Google Agent Compute Engine: a hardware microVM per agent, zero credentials in the sandbox
Google Agent Compute Engine (ACE): Google Cloud industrializes the execution infrastructure of its AI agents. Hardware MicroVM per agent, zero credentials.
Harness²: Google Research lets agents rewrite the system that surrounds them
Harness²: the framework from Google Research that lets AI agents rewrite their own prompts, tools, and memory. The model is no longer the only lever.
OpenGCUA: the open stack for computer-use agents that blasts through OSWorld-v2 with 53% fewer steps
OpenGCUA: the open stack for computer-use agents that saves 53% of steps on OSWorld-v2. The open alternative to proprietary agents.
Anthropic's report on Codex 5.3: 58% of bugs found and failures that declare themselves "successes"
Anthropic's Codex 5.3 report reveals 58% of bugs found but failures declared 'successes'. What this says about the reliability of code agents.
Google launches Gemini Agent: the universal agent that wants to take over the entire workstation
Google launches Gemini agent: the universal agent that wants to take over the entire workstation, with business context and a deliverable in your tools.
Wikipedia denounces OpenAI's "rogue" agents: when AI agents crush the web's infrastructure
The Wikimedia Foundation reveals OpenAI's 'rogue' agents on Wikipedia: when AI agents crush the web's infrastructure.
The "harness" layer is becoming infrastructure: two of the three trending GitHub repos are agent harnesses, not models
Two of the three trending GitHub repos are agent harnesses, not models: the harness layer is becoming key AI infrastructure in 2026.
Cursor opens its ecosystem: the official plugin spec arrives on GitHub trending (4,700 stars in just a few days)
Cursor opens its ecosystem: the official Cursor plugins spec is making waves on GitHub with 4,700 stars in just a few days.
OpenAI launches Dots: always-on personal agents to counter Meta Muse
OpenAI launches Dots, always-on personal agents powered by GPT-6 Astra, to counter Meta Muse. Discover the DevDay 2026 announcements.
Hindsight: AI agent memory becomes a product in its own right (+1,653 stars in one day)
Hindsight, the open source memory layer for AI agents, is blowing up on GitHub with +1,653 stars in one day. AI agent memory is becoming a product.
Laya: the open-source 421M param "decision model" that wants to replace the LLM in your agents
Laya, the open-source 421M-param decision model that wants to replace the LLM for routing, scoring and validating your AI agents' decisions.
**google/ax v0.3.0: Google's agent orchestration runtime goes open source and takes off (+2,305 stars/day)**
google/ax v0.3.0: Google's agent orchestration runtime goes open source and is booming on GitHub with +2,305 stars per day.
Meta Muse Sends the Stock Market Tumbling: Schwab -6.7% and S&P Financials -2% Due to an AI Agent
Meta Muse, Meta's AI agent, sends stocks tumbling: Schwab -6.7%, S&P Financials -2%. Discover how an AI agent shook Wall Street. *(127 characters — within the 160-char limit)*
"Skills" are taking over GitHub trending: obra/superpowers and mattpocock/skills are inventing the new standard for coding agents
Skills GitHub trending: how obra/superpowers and mattpocock/skills are redefining the coding agent standard with simple Markdown files.
Amazon blocks Muse: Meta launches an agent that shops on your behalf, Amazon says "no" within hours
Amazon blocks Muse: Meta's AI agent that shopped on your behalf is cut off within hours. Analysis of the first duel between AI and e-commerce.
**Plugin4Shell: the zero-click RCE that traversed Claude Code, Codex, Copilot, and Gemini CLI — and the two vendors that never patched it**
Plugin4Shell: the zero-click RCE affecting Claude Code, Codex, Copilot and Gemini CLI. Analysis of the flaw and of vendors who never patched it.
AIUC raises $55M to bring SOC 2 to AI agents: 5,000 jailbreak, hallucination, and data leak tests before signing an enterprise contract
AIUC raises $55M to create a SOC 2 for AI agents: 5,000 jailbreak, hallucination and data leak tests before any enterprise contract. *(132 characters — within the 160 limit. No URLs, slugs, or tool names present in the source text.)*
**Meta Muse: OpenClaw becomes a mass-market product — Meta's personal agent arrives on WhatsApp**
Meta launches Muse, its personal AI agent on WhatsApp and mobile: emails, travel, smart home. OpenClaw becomes a mass-market product for 3 billion us
forkd : the Rust project that forks AI agent micro-VMs in 100 ms — agent sandboxing rethought
Discover forkd: the revolutionary Rust project that forks AI agent micro-VMs in 100 ms. Agent sandboxing, reimagined.
OpenViking (ByteDance): the self-evolving context base that unifies memory, RAG, and skills of AI agents - 30,000 stars trending on GitHub
Discover ByteDance's OpenViking: the self-evolving context base unifying AI agents' memory, RAG, and skills. 30k-star GitHub project.
Prime Agent: The #1 GitHub repo (17,500 stars) that makes code agents self-improving — and auditable
Discover Prime Agent, the #1 GitHub repo with 17,500 stars. This self-improving coding agent delivers enhanced performance while remaining entirely a
Agent Skills : the 68k-star repo that teaches AI agents engineering best practices
Discover Agent Skills, the 68k-star repo teaching AI agents engineering best practices to avoid technical debt.
OpenAI launches ChatGPT Work: the AI agent that works for hours without you
Discover ChatGPT Work, OpenAI's new autonomous AI agent that executes complex tasks for hours without any human intervention.
Qwen-AgentWorld : when an LLM simulates the world to train autonomous agents — the new frontier of language world modeling
Discover Alibaba's Qwen-AgentWorld: a revolutionary LLM that simulates the world to train autonomous agents. The new frontier of language world mo
Agentic Resource Discovery: the open standard that will unify AI agents
Discover Agentic Resource Discovery, the new open standard from Google and Microsoft designed to unify AI agents and automate their tool discove
Google launches the Interactions API in general availability: the new default interface for building Gemini agents (and generateContent retires)
Google launches Interactions API to GA. Discover the new default interface for your Gemini agents and the end of generateContent.
Vercel eve: the open source framework that wants to do for AI agents what Next.js did for the web
Discover Vercel eve: the revolutionary open source framework for building production AI agents, just as Next.js transformed the web.
OpenAI Codex Record & Replay : show a task once, the agent repeats it endlessly — the end of manual scripting
Discover OpenAI Codex Record & Replay: show a task once, and the agent repeats it infinitely. The end of manual scripting is here.
Claude Code switches to monthly credits: what changes for devs and autonomous agents
Anthropic ends Claude Code's free tier. Discover how the new monthly credit billing changes things for developers and autonomous agents.
MiroFish: an undergrad builds 700,000 AI agents in 10 days — this open source project predicts the future and explodes on GitHub
MiroFish: a student creates 700K AI agents in 10 days. Discover this open source project exploding on GitHub and predicting the future.
EEVEE : the first test-time prompt learning framework for self-improving AI agents
Discover EEVEE, the first test-time prompt learning framework designed to create self-improving and adaptable AI agents in real time.
Life-Harness : boosting LLM agents by 88.5% without retraining, the open source runtime revolution
Discover Life-Harness: the open source runtime revolution boosting LLM agents by 88.5% without retraining. No more brute force!
OmniGameArena: The UE5 benchmark that measures the learning dynamics of VLM agents in games
Discover OmniGameArena, the UE5 benchmark to evaluate VLM agents' learning dynamics in video games, beyond mere scores.
Life-Harness : boosting LLM agents by 88.5% without touching the model — the runtime revolution
Discover Life-Harness: the runtime method boosting LLM agents by 88.5% without fine-tuning. No more production failures!
OmniGameArena : the UE5 benchmark revolutionizing the evaluation of VLM agents in games
Discover OmniGameArena, the revolutionary UE5 benchmark for evaluating VLM agents in games. Forget simple scores and measure real progression.
StreamMA : Streaming that reduces multi-agent latency by 50% — a new paradigm for distributed reasoning
Discover StreamMA: a streaming approach that cuts multi-agent latency by 50%. The new paradigm of distributed reasoning in 2026.
Robinhood Agentic Trading: AI trades stocks on your behalf — and it's changing finance
Discover Robinhood Agentic Trading: an autonomous AI that trades stocks for you for 27 million clients. Finance is changing.
Unlocking Working Memory: this research shows how LLMs can reason without generating tokens
Discover how LLMs can now reason without generating tokens. Research revolutionizing how language models work.
FluxMem : when AI agent memory learns to evolve like a brain
Discover FluxMem: AI agent memory that evolves like a brain. Solve the forgetting problem in your complex workflows.
Devin Auto-Triage: Cognition transforms its AI agent into a 24/7 on-call engineer
Discover Cognition's Devin Auto-Triage: an AI agent turned 24/7 on-call engineer. No more on-demand mode, the AI manages your production incidents.