We've been building quite a few open-source tools at Dailybot lately, and I want

By Maomorales · October 20, 2026 · Curated by George's Blog

We've been building quite a few open-source tools at Dailybot lately, and I wanted to share some of what we're working on.

Our engineering team has been experimenting a lot with AI agents, particularly around how to make them more autonomous, coordinate multiple agents working in parallel, and how to make all of this actually useful in real development workflows.

As we've been experimenting, we've also been building our own tools to solve some of the problems we've encountered. Some of these projects have evolved quite a bit, and we've decided to make them open source so others can use them, contribute, or build on top of them.

A few things we've been working on:

🧠 Deep Work Plan: A methodology and set of skills for AI agents to plan, execute, and validate complex development tasks

🤝 Herdr Peers: A way for multiple AI agents to communicate and coordinate work across machines

⚡ Coding Agents Kit: Run different coding agents, including Claude, Codex, Grok, and others, from one toolkit

🔍 AI Diff Reviewer: Automated code reviews using AI, locally or through GitHub Actions

🐳 Dev Container Kit: Development environments ready for coding agents to work in

💻 DWP Vim: A terminal-based editor with visibility into Deep Work Plans

🤖 Dailybot Agent Skill: Connect AI agents with Dailybot so they can report progress, interact with teams, and communicate through Slack, Teams, Discord, and other platforms

These are separate projects, but they also work together as an ecosystem.

We're particularly interested in what happens when you move beyond using gone AI agent at a time, and start having multiple agents collaborating, executing longer tasks, and working with less (or none) human intervention.

We're learning quite a bit by building these tools and using them ourselves, and you can explore and try this, everything is in the following GitHub page.

View the original post on LinkedIn