Build A Wall Street Analyst Inside Claude

Anthropic open sourced its own financial-services skills: production-grade agents and slash commands that draft the exact work product a junior analyst produces. DCF valuations, LBO models, comps, three-statement models, IC memos, and pitch decks, as real Excel and PowerPoint files, from one prompt. Here is the full setup, end to end.

What You'll Learn

• The one-line install for Anthropic's official financial-services plugin
• The exact slash commands: /dcf, /comps, /lbo, /ic-memo
• The honest part: bring your own numbers, it builds the model

What It Actually Is

Anthropic open-sourced its own financial-services skills on GitHub (anthropics/financial-services, Apache 2.0, 32,000+ stars). It is a set of production-grade agents and slash commands that draft the exact work product a junior analyst produces: DCF valuations, LBO models, comps, three-statement models, IC memos, and pitch decks, as real Excel and PowerPoint files.

These are the same modeling workflows banks and hedge funds pay a fortune to run. A single Bloomberg Terminal seat is $31,980 a year, and the full analyst software stack runs into six figures. Anthropic gave away the analyst's brain for free. You still bring the data, and a human signs off, but the workflow is yours.

One-Time Setup

Add Anthropic's official marketplace inside Claude Code (or Claude Cowork), then install the finance plugin bundle. Restart Claude afterward so the slash commands load.

/plugin marketplace add anthropics/financial-services
/plugin install financial-analysis

The repo ships several bundles (investment banking, equity research, private equity, wealth, fund admin), so run /plugin to see them and install the one you want.

The repo: https://github.com/anthropics/financial-services (Anthropic's official open-source financial-services skills, Apache 2.0, free, 32,000+ stars).

Build A DCF In One Prompt

This is the headline move. Building a discounted cash flow model by hand is an afternoon of work for an analyst. Here you type one command, point it at a company, and it builds the whole model as a real Excel file: live formulas, the valuation, and a sensitivity table.

/dcf Build a DCF for Apple. Use the latest annual figures I paste below, project 5
years, and include a WACC sensitivity table.

Bring your own numbers: the premium data connectors (FactSet, PitchBook, Bloomberg) still cost money. You do not need them. Paste the figures straight from a company's public 10-K or annual report and it builds the model around them. The skill drafts, you sign off. It does not place trades or post anything anywhere.

The Rest Of The Analyst's Desk

The same plugin runs the rest of the workflow. Each command drafts a real, formatted deliverable you then review and edit:

• /dcf: a discounted cash flow model as a live-formula Excel file with a sensitivity table
• /comps: a comparable-companies analysis
• /lbo: a leveraged buyout model
• /ic-memo: the investment committee memo

The memo that saves the night: /ic-memo writes the investment memo a junior analyst would lose a whole night to. It pulls your model, the comps, and the thesis into one formatted document, staged for your sign-off.

Where To Go From Here

Modeling a single company is the easy win. The real leverage is that you now have an analyst on call for any decision with numbers in it: pricing a product, valuing a deal, sizing a market, reading a competitor's filing. The workflow that used to live behind a six-figure software bill now runs from one prompt.

Nothing here is financial advice. The skills draft models and memos from numbers you supply; every output needs a human review before any real decision.

Go Deeper

• Printable PDF version of this guide: https://drive.google.com/file/d/1fwuS2LB2e2UbAG59cr8OG7J7GNuFAefU/view?usp=drivesdk

Want the whole system installed live, with Cooper on the call? Build Your Own AI, the two week sprint runs the first two weeks of August.

Or get a new working tool every Monday inside The AI Operator community on Whop.