A Free Claude Skill That Knows Every Portfolio123 Factor

We are releasing the Portfolio123 Claude Skill v3.0.0 as a free download for the whole quant community. It covers every factor, every function, and the full API, and all of it has been verified against the official documentation.

Download portfolio123.skill (free)

If you have ever asked an AI assistant to write a Portfolio123 screen, you already know the problem. It invents factor names that do not exist. You ask for market cap and get MarketCap instead of MktCap. You ask for the Piotroski score and get PiotroskiF instead of PiotFScore. Or it produces ranking-system XML that P123 rejects on upload, and you end up spending more time fixing the AI’s output than it would have taken to write the rules yourself.

This skill exists to fix that. It hands Claude the complete vocabulary of the Portfolio123 platform, checked name by name, so the screens, ranking systems, and API code it writes work on the first try.

What’s inside

Nothing in this skill comes from an AI’s memory. Everything was extracted programmatically from P123’s official Factor Reference and API documentation, then validated entry by entry. The package includes:

All 4,463 factors and 465 functions across the 13 official categories (Ratios & Statistics, Financials, Fundamentals, Estimates, Technical, Advanced Functions, Strategy, and the rest), each category in its own reference file with official descriptions and period variants.
473 extras that most references skip: formula constants, macro series IDs with their FRED mappings, country and universe IDs, and the formula-language operators.
The full REST API. All 33 operations, the complete 38-method map of the p123api Python wrapper, authentication and quota semantics, and a Known Pitfalls section for the places where the spec and the real API disagree.
The validated ranking-system XML schema, including a worked multi-factor example and a table of known formula errors to avoid.
Nine ready-to-run Python scripts covering auth checks, screen runs, backtests, rank downloads, bulk data pulls, price history, and AI Factor predictions. Every script is read-only by default. The one script that can change your account is locked behind an explicit --execute flag.
A “Common Mistakes” table in every file: the names AIs love to invent, listed right next to the verified correct ones.

How to install it

In Claude (claude.ai and the desktop app, including Chat and Cowork)

Download portfolio123.skill with the button above.
In Claude, open Settings, go to Capabilities, then Skills, and click “Upload skill”.
Select the downloaded portfolio123.skill file.

From then on the skill activates by itself in your chats and Cowork sessions whenever Portfolio123 comes up. Note that skills require a paid plan with code execution enabled.

In Claude Code (CLI and IDE)

Download and unzip portfolio123.skill. It is a standard zip archive.
Move the portfolio123 folder into your skills directory: ~/.claude/skills/portfolio123 to have it in all projects, or .claude/skills/portfolio123 inside a single project.
Start a new Claude Code session and ask anything P123 related. The skill loads on its own.

What to ask once it’s installed

“Write a P123 screen for profitable small caps under 15x earnings with improving margins.”
“Build a 4-node value/quality ranking system as XML.”
“Why does my rule Eval(IsNA(PEExclXorTTM), ...) fail on P123?”
“Download PE, ROE, and 12-month momentum for the SP500 universe as a pandas DataFrame.”
“Which P123 factor is the Piotroski score, and how do I rank on it?”

How we checked the work

We held this release to the same standard we hold our own research. An automated checker validated every factor name in every file against the extracted dictionary. Key read-only API calls (authentication, a screen run, and a price-history download) were exercised live against Portfolio123’s production API before shipping, and everything else was cross-checked statically against the official OpenAPI spec and the wrapper source code. On top of that, an independent verification pass re-sampled entries against the live documentation pages, and a second adversarial round re-verified every fix before we called it done. The build log ships inside the package, so you can read exactly what was checked and how.

The skill is MIT licensed and open source. You can browse the code, open issues, or contribute on GitHub. It is an independent community project, not affiliated with or endorsed by Portfolio123, Inc.

Get the skill

Give Claude the real Portfolio123 dictionary and spend your time on research instead of proofreading factor names.