For coding agents, "skills" are a great way to automate repetitive workflows, but how can we tell if they're working at scale? We did a deep dive on how you can log, monitor, and improve agent skills, with a real example of building a customized PR review skill.

The team is at @aiDotEngineer World's Fair this week in SF! If you’re attending and want to talk about: 🔁 Automating repetitive engineering toil: CVE remediation, refactors, or dependency upgrades 🔒 Running autonomous agents safely in sandboxes across your codebase 🏠 Self-hosted, VPC, or on-prem deployments for full data control 📈 Scaling from local experiments to thousands of parallel agent jobs 👋 Send us a note or stop by the booth. We'd love to connect. #AIEWF #AIEngineer #AIWorldsFair #AgenticEngineering

We've found this sort of "sidekick" architecture to be very effective at cutting LLM spend because it allows you to do context control and not spend expensive tokens on simple tasks. Here's a 200-line example of how to do it in the OpenHands SDK :) gist.github.com/neubig/412ab…

The longer the task, the more orchestration carries it. The OpenHands SDK just added three primitives for building robust, long-horizon agent workflows, usable by the model and by you: Thanks to: @gneubig @juanontech Vasco Alona @OpenHandsDev

OpenHands listed as a top #opensource project for #AI-driven development 🙌

Hot off the press 📰🔥📰🔥 "As organizations increasingly adopt agents for developer productivity, code automation, and other engineering tasks, many teams are hitting a scale issue as agents sprawl across repositories, teams, and business units. OpenHands’ offering responds to that challenge by treating agents not as independent, one-off tools but as components of a coordinated system, managed from a single operational plane." aijourn.com/openhands-launch…

We open-sourced a method to cut slop from agent-generated code. It seriously works, and we have receipts -- PRs on production code merge 2.4 times faster while still maintaining quality. See Xingyao's post below for details on *The Verification Stack*

Coding agents open PRs in minutes. But on consequential codebases, those PRs still wait days for human review. So we built a verification stack, and on our own repo, it cut time-to-merge by 58% without lowering code quality. Here’s how. 🧵

Another day, another benchmark that @OpenHandsDev tops 😄

OpenHands is quietly becoming the default open source AI coding agent everyone’s actually using.,, 76k stars in a few months!! Its a full autonomous agent that can plan, code, debug, and push changes by itself. What’s making it stick :: --> Works in terminal or GUI so you can literally watch it work --> Supports Claude, GPT, and local models without lock-in --> Handles real tasks end-to-end instead of just spitting out code snippets --> Actively maintained with frequent updates and a growing builder community Most agent tools still feel like demos,, But openhands feels like something you can actually run on real projects without babysitting it every five minutes. If you’re building or testing AI agents for development work right now, this is currently one of the strongest open options available.

If you’re building AI agents, study these projects: • @polygres — Postgres for the agent era • @LangChain / LangGraph — stateful agents • @llama_index — agents over your data • @crewAIInc — multi-agent workflows • @dify_ai — agentic app builder • @FlowiseAI — visual agent flows • @OpenHandsDev — coding agents • @browser_use — browser agents agents need memory, tools, retrieval, workflows, and data that actually makes sense.

People have been asking why OpenHands V1 goes the opposite direction from Claude Managed Agents. I finally found the time to write it up. The production pain points are real, but changing the hosted agent architecture won't solve them for free: xwang.dev/blog/2026/openhand…