The week in brief
Behind it, OpenHands (+577) and GPT Researcher (+147) followed.
On Hacker News, “We ran AI agent swarms in Minecraft and Mindustry” drew the most attention in this space (421 points, 174 comments). Among listed tools, Claude's site was linked from 45 Hacker News stories over the past 90 days, more than any other tool here.
In research, “Wave Forcing: Towards Speed-of-Light Streaming Video Generation” collected the most upvotes on Hugging Face Daily Papers this week (82). Among new GitHub repositories created in the past 7 days, wangzan101/Mika leads with 5,835 stars.
Biggest GitHub star gainers
GitHub star counts for the 16 open source tools we track, compared between Oct 27, 2026 and Nov 3, 2026 (169 hours apart).
| Tool | Stars before | Stars after | Gain | Change |
|---|---|---|---|---|
| Browser Use | 119,262 | 119,915 | +653 | 0.55% |
| OpenHands | 91,983 | 92,560 | +577 | 0.63% |
| GPT Researcher | 30,313 | 30,460 | +147 | 0.48% |
| garak | 9,721 | 9,842 | +121 | 1.24% |
| AutoGPT | 187,785 | 187,886 | +101 | 0.05% |
| DeepSeek | 104,628 | 104,722 | +94 | 0.09% |
| STORM | 31,707 | 31,783 | +76 | 0.24% |
| Inspect | 3,109 | 3,157 | +48 | 1.54% |
Most discussed on Hacker News
The highest scoring Hacker News stories in this space during the week.
Listed tools with the most Hacker News links
Hacker News stories linking to each tool's own website over the past 90 days.
Notable new papers
The most upvoted papers on Hugging Face Daily Papers and the newest relevant arXiv preprints from this week.
- Wave Forcing: Towards Speed-of-Light Streaming Video Generation
- Video Generation Meets Agent Harness: Agentic Video Generation, Agentic World Modeling and Recursive Self Improvement
- CoworkOS:Scaling Agent Coworkers for Real world Automation
- CUA-Master: Scaling Computer use Agents through Open World Recursive Improvements
- Democratizing Robot-Use Agents: A Unified Harness with Open Models
- WebHarbor: Docking Real Websites for Evolving GUI Agent Environments
- Strategist: Evolving Reasoning with Agents
- DiabAgent: Toward Agentic Physiological Control
- Four Bridges: Measuring Deception Propensity in a Multi Agent Game
- Spacetime Emergence in the Memory of a Learning Agent
New GitHub repos in LLM
Repositories created in the 7 days before Nov 03, 2026, ranked by stars.
| Repository | Stars | Language | Description |
|---|---|---|---|
| wangzan101/Mika | 5,835 | Kotlin | An on device AI agent for Android that operates apps and the web to get things done for you. |
| N1nEmAn/sapper-v2 | 652 | Python | Autonomous pentest legion. AI commander + worker swarm + layered fact intent graph. 3rd Tencent Intelligent Attack-Defense online champion (634 teams · 8h unattended · 1B tokens beat 4.5B). The Nüwa |
| Balanov/uncut | 580 | Go | The lossless, headless coding agent harness. One 7 MB static binary, four tools, plain stdin/stdout, no silent cuts. |
| wanghao-SunSky/cs329z-chinese | 476 | HTML | 斯坦福 Stanford CS329Z Engineering AI Agents 中文教程：面向 Python 开发者的 AI 智能体课程，19 讲与 4 项实践，涵盖 LLM、RAG、工具调用、记忆、多智能体、评测与安全。完整 Markdown、在线与离线 HTML。Chinese course notes. |
| jmanico/grim-prompter | 350 | Threat modeling for AI coding agents: point Claude Code, Codex or any agent at your repo and get an evidence traced threat model. 36 architectures, 29 methods, MIT. | |
| aa1143/NamiAgent | 268 | Python | 零基础入门 AI Agent 的基础教程。不只是教程，也是一个能在本地跑起来的 Agent：读写文件、执行命令、MCP、长期记忆，自带网页端看清每一步。从 100 行最小循环开始，一课加一个能力，最后拆解 Claude Code、Codex 等主流 Agent |
| muaaz9329/cheap-ag | 241 | TypeScript | 50x your usage with cheap ag. Opus/Astra plans, a dirt cheap model writes every line, and tests decide what ships. Your $20 plan stops hitting limits. One MCP server for Claude Code & Codex. |
| SAOS-3624/dsh-see-now | 224 | JavaScript | 让 AI 实时看见你的屏幕：按变化量筛选帧、多帧拼进一张图，静止画面几乎不花 token |
How we compile this report
Every figure comes from a public source: star counts from the GitHub API, story scores and links from Hacker News (via Algolia), papers from arXiv and Hugging Face Daily Papers, and new repositories from GitHub search. Hacker News link counts include stories whose link points to the tool's own domain. Listing dates are the day a tool joined Intelsi. Numbers are shown as recorded; if a figure is unavailable, it is left out rather than estimated.