Agent on Lynx Tech Blog

Doubao Work: The Agent Skips Re-Onboarding and Reads Your Feishu Org Context Directly

Sat, 29 Aug 2026 08:00:00 +0800

Every mainstream office Agent can do the work — process files, generate web pages, build spreadsheets — but nearly all of them first need you to “explain the background.” ByteDance’s Doubao Work (released August 26, 2026, covered by QbitAI, TechNode, and Caixin) flips that around: log in with your Feishu enterprise account, and it reads your group chats and cloud docs directly, no re-explaining required. What it’s going after is the “zero-config context access&rdquo

Installing a Control Plane for 'Script + Cron + Agent' Black-Box Automation: A Survey of 12 Platforms

Sun, 23 Aug 2026 00:00:00 +0800

TL;DR: What I need isn’t another orchestration tool, but a Control Plane—a coordination layer that answers in one place: which workflow, which version, which run, which step failed, how many retries, when’s the next run. cron only knows how to “knock on time” and then walks away. After surveying 11 platforms, my pick: Windmill first choice, Kestra co-selected; n8n/Temporal/Argo explicitly rejected; Hermes demoted to “scheduled agent” instead of scheduler.

5 Days, 150K Stars: DeepSeek Harness Sells Agents à La Carte

Wed, 19 Aug 2026 08:00:00 +0800

On the evening of August 13, DeepSeek dropped this year’s most significant open-source move—DeepSeek Harness (DSH). This isn’t just another large model update; it’s DeepSeek’s first time open-sourcing its own AI Agent runtime framework, released under the MIT license and free for commercial use. Five days later, GitHub stars surged past 150,000 with 15,000 forks, shooting it into the ranks of GitHub’s fastest-rising projects of all time. For comparison: DeepSeek&rsq

Prompts, Context, Harness: How the Focus in AI Has Shifted Three Times Over the Past Four Years

Mon, 10 Aug 2026 04:30:00 +0800

First, an Experiment In early 2026, developer Can Bölük ran an experiment: same model, same set of tasks, nothing changed except the engineering layer around the model—specifically, the format used by the harness to handle code patches. The task success rate jumped from 6.7% to 68.3%. Tenfold. Not a single line of the model changed. That number has been making the rounds in AI circles lately because it turns something many people had vaguely sensed into a conclusion that is impossible to ignore:

Hermes vs OpenClaw: Which Self-Hosted AI Agent Harness in 2026

Fri, 31 Jul 2026 00:00:00 +0800

The two rising self-hosted agent harnesses Hermes and OpenClaw are two open-source, self-hostable AI agent harnesses attracting growing attention in 2026. Both let you run an agent that talks to humans across messaging channels, calls tools, and maintains state — but they differ in language, philosophy, and ecosystem. Dimension Hermes OpenClaw Repo NousResearch/hermes-agent openclaw/openclaw Version (2026-07-31) 0.16.0 2026.7.1-2 Language Python Node.js / TypeScript License MIT MIT One-liner The

Why I self-host my AI agents (and built AgentHub)

Fri, 31 Jul 2026 00:00:00 +0800

A self-hosted rack of agents reaching messaging channels|AI-generated illustration Every day I run two self-hosted AI agent harnesses: Hermes (from Nous Research — Python, self-improving, with a skills system) and OpenClaw (a Node-based multi-channel agent gateway with a skills marketplace). Both are MIT-licensed, both live on my own machine, and their channels cover everything from WeChat, Feishu, and DingTalk to Telegram.