Use Claude Code for twice-weekly stock research

Claude Code can track a portfolio, compare bullish and bearish views, and leave trade approval to a human.

Done with Claude Code· Intermediate

What they did
An anonymous operator told Claude Code, in one short instruction, to try to outperform the S&P 500 in a live $50,000 account. The agent was then used as a research helper for about 15 holdings and a separate list of candidates, pulling together fresh headlines, scenario analysis, and opposing views twice a week. The operator defined the guardrails: position-size caps, a limit on trading turnover, and a requirement to write down a thesis that could be disproved; they say the final selections were theirs.
How it went
The operator reported a gain of 19.04% between March 3 and August 4, 2026, compared with 12.24% for the S&P 500, and said $27 million was following the trades. The article also said the figures were not audited, five months is too little time to tell whether the edge is real, and the setup functions as part of a copy-trading sales pitch.
Tips
Keep the agent on the research side and make every trade wait for human review. Spell out risk rules before you start, and remember that short-term outperformance may not mean the system works.

Try it yourself with Claude Code

Act as my research analyst for [holding list] and [watchlist]. Twice a week, check recent headlines for each name, summarize the key points, outline reasons to like it and reasons to avoid it, and call out anything that would violate these rules: [risk rules]. If a change, order, sale, deletion, or submission might follow, stop and ask for my approval first. Deliver the first report in [format].

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

Read the original on explainx.ai ↗

Get 3 agent uses like this every morning

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