Claude Code read a directory of brokerage CSV exports and notes, researched funds, and produced buy, stop-buying and hold lists with a tax-loss harvesting partner fund for each.
Claude Code read a directory of brokerage CSV exports and notes, researched funds, and produced buy, stop-buying and hold lists with a tax-loss harvesting partner fund for each.
Matt Stockton pointed Claude Code at a directory with brokerage CSV exports, a typed scratchpad of other balances, and a detailed goal prompt covering constraints, strawman targets and "ask me questions exhaustively". Claude wrote Python to parse and classify holdings, then built a gap analysis. Over several sessions he refined constraints, and a CLAUDE.md file carried context between them.
It produced a ten-section plan with a four-phase action plan, tax payback tables and fund research. He's nearly done executing it. A self-rating of 7.5/10 surfaced improvements, such as using tax-free accounts for the overweight position.
The goal prompt mattered most. Include constraints, such as accounts you won't touch, and ask Claude to critique its own plan. It took a few evenings across multiple sessions.
Read the brokerage CSV exports and notes in [folder path]. Research the funds I hold and give me three lists: buy more, stop buying, and hold. For each fund you suggest selling at a loss, name a similar partner fund for tax-loss harvesting and explain the reason. Write the result to a file called portfolio-plan and don't make any trades.
An app built with Claude Code extracts vendor, date, total, tax and line items from receipts, including foreign-currency ones at historical exchange rates, for freelancers and small businesses.
A tenant used ChatGPT and Gemini to review a landlord's apartment deposit dispute, identify legal violations, and avoid losing $4,200.
The poster used an AI assistant to summarize a denial PDF to draft an appeal, but it invented an appeal deadline.