Skip to main content

​
What is Idea Screening?

Stock screening is the process of filtering the universe of publicly traded companies down to a shortlist of potential investment ideas using quantitative criteria. Instead of analyzing thousands of stocks one by one, you set filters — such as “P/E below 15x, revenue growth above 10%, and insider buying in the last 90 days” — and the screen returns companies that meet all your criteria. Screening is how many of the best investment ideas begin. Quantitative hedge funds like Renaissance Technologies and Two Sigma run screens algorithmically, but even fundamental investors at firms like T. Rowe Price and Capital Group use screens as a starting point before doing deep qualitative research. The key insight is that screens surface candidates, not conclusions — every screen output needs fundamental analysis before it becomes an investable idea.
The most interesting ideas often come from intersections of screens — for example, a high-quality company (consistent growth, high returns) trading at a value price (below-average multiple) due to a temporary headwind. These situations arise when the market over-reacts to short-term problems at fundamentally strong businesses.

​
Command Syntax

/screen [screen criteria, e.g. 'undervalued midcap tech']
If criteria are not provided, the command asks about direction (long/short), market cap, sector, style (value/growth/quality), geography, and theme.

​
What It Produces

  • Shortlist of 5-10 ideas with one-page summaries per idea
  • Screening criteria and methodology documented
  • Comparison table across all ideas
  • Prioritized list of which ideas to research further

​
How to Customize

  • To add proprietary screening factors or data sources, edit the idea-generation skill file
  • To connect to financial data APIs for live screening, configure them in .mcp.json