Orchestration and shared state, in brief

About this track: Clawdemy uses the open-source TradingAgents framework (by TauricResearch; Yijia Xiao, Edward Sun, Di Luo, and Wei Wang; arXiv:2412.20138; Apache-2.0) as a real-world example of how an agentic AI system is structured. Our goal is to teach the architecture: how specialist agents, a research debate, a trader role, and a risk layer are coordinated by an orchestrator. We are not teaching you how to invest, trade, or pick stocks, and we make no claim that this or any AI system is profitable or predicts markets. This content is for education only and is not investment, financial, or trading advice. Rules vary by country, and this is not advice anywhere. The Trading Agents Lab capstone is a simulation you explore yourself with your own AI provider key; it does not place real trades. The framework’s own authors state it plainly: “It is not intended as financial, investment, or trading advice.”

What you’ll learn

The first five lessons walked the team role by role. This one zooms out to the wiring that makes those roles a single system, and it turns out to be two separate things. Control flow is the question of who runs next: some handoffs are fixed (after the judge, always the trader), and some are conditional (loop the debate again, or move on to the judge). Data flow is the question of how one agent’s work reaches the next: the whole team shares one structured workspace, a single source of truth that every agent reads from and writes to. Reading the real code, you will see the handoffs from earlier lessons reappear as plain reads and writes of named spaces on that shared board, and you will see the “should we continue” checks from the data-fetching, debate, and risk lessons revealed as the conditional edges of one graph. The throughline is transferable: keep one source of truth, and be deliberate about which handoffs are fixed and which adapt.

Where this fits

This is the sixth lesson of the track. Lessons 1 through 5 introduced the team and followed a decision from gathering through debate, judgment, planning, and the risk gate. This lesson is the step back that shows what was connecting all of them the whole time. Next, lesson 7 looks at memory: how the system records its past decisions and reflects on how they turned out, and the track finishes with the hands-on capstone.

Before you start

Prerequisites: lesson 5 of this track (the risk gate). It also helps to have the earlier lessons fresh, because this lesson ties their handoffs together. You do not need to install or run anything; the project can be read in your browser, and every excerpt is quoted in the lesson.

By the end, you’ll be able to

Distinguish control flow (who runs next) from data flow (what agents share)
Explain how a single shared state acts as the team’s source of truth
Differentiate a fixed handoff from a conditional one, and know when each fits
Recognize the earlier “should we continue” checks as conditional edges of one graph
Apply the source-of-truth and fixed-versus-conditional patterns to your own AI workflows

Time and difficulty

Read time: about 10 minutes
Practice time: about 15 minutes (a self-check, a design exercise, and flashcards)
Difficulty: deep (this track is the advanced end of the catalog; the foundational agent tracks are the on-ramp)