Screen Fictional Companies With Claude Code Equity Research
Build a reproducible company screen from explicit eligibility rules and keep the observed comparables separate from investment conclusions.
中文Setup for Claude Code
Claude Code and the equity-research plugin. A small local CSV is enough for this exercise. Live universe and research data are separate prerequisites.
- In Claude Code, run claude plugin marketplace add anthropics/financial-services.
- Install with claude plugin install financial-analysis@claude-for-financial-services.
- Add the workflow with claude plugin install equity-research@claude-for-financial-services.
- Start Claude Code in a scratch directory. Invoke /screen or ask for the idea-generation skill by name.
Example task
Create a fictional CSV of Atlas, Birch, and Cedar with revenue growth of 12, 8, and 14 percent and net margins of 15, 20, and 6 percent. Ask for growth greater than 10 and margin at least 10 percent, plus an exclusion reason for each failed row.Expected result
Only Atlas passes. Birch misses growth and Cedar misses margin. The result lists the thresholds and preserves the figures used for the filter.
Upstream documentation
Original SKILL.md on GitHub · Apache-2.0 license · Source reviewed
Frequently asked question
How can I make a screening run reproducible?
Save the input table and exact comparison operators, then record which rule excluded each row.