Build a reusable stock-analysis workspace

A custom agent kept one investment project alive, saving models and notes so later sessions could pick up where the last run stopped.

Done with Custom agent· Advanced

What they did
They started by assigning the agent a single investing task and letting it ask questions about the goal and preferred analysis style. The agent then worked in a Daytona sandbox to collect financial data and generate outputs such as DCF and comps models, while writing its files and a memory log inside the project space. It also split smaller jobs across subagents, and they could adjust the direction during the run.
How it went
In one example, the agent produced an Excel comps file and a PDF summary showing estimated share-value ranges for NVDA, AMD, and GOOGL. Because the workspace and notes were preserved, a later session could reopen the project and continue the analysis. The README did not include any quantified performance outcome.
Tips
This setup is most useful when you keep one workspace tied to one research question and capture assumptions as you go. If you self-host, Docker and your own LLM key are enough to run it, but the data feed is thinner without extra market-data keys and some automations stay on the hosted service.

Try it yourself with Claude Code

Set up a continuing research project for [company or ticker] in [folder]. Start by asking me what I’m trying to learn, what peers to compare against, and what kind of valuation output I want. Then pull the newest filings, market prices, and relevant headlines, and create a simple model plus a peer comparison table. Save assumptions and open questions in a memory file so the next session can resume from there. Before any action that would send, submit, delete, or pay for something, pause and ask me to approve it. When you’re done, show me the first model and the saved notes.

Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.

Read the original on github.com ↗

Get 3 agent uses like this every morning

With the day's top AI news. Free, and one click to unsubscribe.