Agent OS 2.0 Progress Update

[CasJam]

Hey everyone! I wanted to share a quick update on Agent OS 2.0 development.

First, a huge thank you for your continued support and feedback – it's been incredibly valuable and motivating! As a solo founder balancing development, content creation, and user support, your patience means a lot.

Timeline

Target release: Early October (subject to change)

What's driving 2.0?

The core mission stays the same – making your agents smarter and more aligned with how you build, plus providing workflow structure for spec-driven development.

The file architecture is getting a significant overhaul. A few factors are driving this update:

• The introduction of Subagents and the efficiency they bring to the goals of Agent OS
• Finding better ways to provide focused instructions without overloading the agent's context
• Building for flexibility to allow for multiple profiles and changing up your workflows and standards across projects, teams, and specs.

Will there be documentation for updating and useage?

Of course :) Agent OS 2.0 will have all new documentation that covers installation, updating, and useage. The docs site will receive a big overhaul to make it easier to navigate and link to specific parts of the new docs.

Migration will be straightforward with clear update instructions for both base installations and existing projects.

Two Modes for Different Tools

Since subagents offer major improvements for process adherence and standards compliance, 2.0 will support:

• Subagent Mode (Claude Code) – Full subagent workflows and orchestration
• Single-Agent Mode (Other CLIs) – Same standards and workflows, single-agent context

I expect more CLIs to add subagent support soon, and the documentation will cover how to adapt Agent OS to use with any tool.

New Features in 2.0

Profiles – Enhanced "project types" with inheritable standards and custom agent/command sets

Spec Research – Kick off a new spec with a round of strategic planning with the agent to better inform the specs and tasks that get created.

Visuals – Ability to add visuals (mockups, wireframes, screenshots) and have agents refer to these for improved UI work.

Playwright integration – Install Playwright and Agent OS workflows can use it to check work in browsers.

Custom Agents – Create your own specialized subagents for specific workflows

Smart Orchestration – Different features automatically get different specialist combinations and workflows

Granular Standards – More targeted organization allowing precise injection into relevant agents

Contributing to v1.x

Bug fixes are always welcome! However, I won't be merging functionality changes since 2.0 has a completely different architecture.

Thanks again for being part of the Agent OS community – excited to get 2.0 into your hands soon!