This is PART 1 of a 2-part series on this topic. The first video covers building the "innermost loop" of the Lobster-like agentic system. It just takes a task definition in English (and a description of the data scheme) and then executes the CFL action formula that will carry out the task. The second part will cover building the complete CodaClaw agent that works just like OpenClaw or Claude Code, but uses Coda Pages and Tables instead of the file-folder tree structure. This allows us to create and manage our Lobsters inside the Coda platform. (They can still take real-world actions through Coda Packs).

All is well at home again, thanks for the good wishes. I have recorded the next workshop and will edit it for better brevity and focus. Then I will post it to the classroom tomorrow. I covers the inner workings of the OpenClaw-like agentic agents that are exciting the AI world right now - and outlines my techniques for recreating one in Coda (my CodaYoda and CodaClaw projects). Will update members by email when it is posted. Max

WRKSHOP 5: AI Agents 🦞CodaClaw & 😈CodaYoda Uploaded I have uploaded the recording to the CLASSROOM and on YouTube

BOOM! 💥 The AI Tsunami has hit ME! 🏄 My clients have cancelled most of my existing projects to pivot in a new direction. All due to the explosion of interest in🦀OpenClaw and fellow Lobster-Like agentic systems! Everyone wants to implement🦀OpenClaw NOW to start getting experience with the business processes it can automate. But all my clients are in regulated industries, and so OpenCLaw is NOT a tool they can deploy. It's just not safe. Especially in the hands of people who are not comfortable with bash commands, GitHub PRs, CLI command windows, VS editing of docs and code. AND because these Lobster-like tools all use the PC hard drive to store instructions, materials, memory, and skills. This is considered unsafe in most controlled environments where all business information must be stored on servers or the cloud. So I have spun up a ton of new client projects where we are using🦐Claude Cowork to explore this new world of Autonomous AI Agentic Workflows - and we are learning what the potential uses are, but in a safe sandboxed environment that still has access to the corporate databases and knowledge bases. As a result of this experience, I have been working out how to replicate the🦞Lobster-like behaviors inside Coda. (Notion and Airtable have already released similar functionality). The result is a series of new projects that I will be delivering here over the coming weeks. (1) 😈CODA YODA: Replicate the innermost loop of OpenClaw/ClaudeCode/ClaudeCowork🦞, etc. A simple loop where you enter a prompt, and it builds the CFL Action formula to implement your request and updates your Coda Pages, Tables, Control-values, etc. Just a proof of concept - but essential for what is to follow (2) 🦞CODA CLAW: Replicate the OpenClaw and ClaudeCowork process inside Coda. The user builds a hierarchy of Coda Pages that contain the skills, references, assets, and instructions (similar to the folder hierarchy we use in OpenClaw) that will direct the behavior of the CodaClaw engine. It will build a strategy and then cycle through the steps to implement that. Each step generates the CFL Action code to cause the changes needed to your Tables, Pages, ControlValues, etc. It can also invoke Coda Packs to interface with the outside world. The loop is autonomous and self-directed. When it finishes, it will notify the user. It is triggered by external events like a webhook or incoming message/email. At first, it will not be able to create tables or columns, but that will be added via MCP later.