Skills

Short answer: Skills are the methodology layer of Questflow’s financial harness — portable judgment systems the AI Finance Agent can load. They cover classic trading strategies (trend, breakout, VCP, turtles) and named investor styles (Buffett, Druckenmiller, Livermore, Ansem, 段永平, and more). Open-source skills live in Investor Skills. Questflow also runs exclusive skills distilled from live fund-manager trades.

What a skill is

A skill is not a tweet-sized tip. It is an operational system: when to use it, what to watch, how to size, what invalidates, what to ignore. In the harness it sits next to models, plugins, and execution.

Questflow exclusive skills go further: they are labeled from real trades — what the manager watched, why they entered, how they sized, what would kill the thesis.

Strategy families the harness is built to run

Family	What the agent is doing	Example skills
Trend / breakout	Follow strength, cut losers, pyramid winners	Livermore, Darvas Box, Turtles, Seykota, Minervini VCP, O’Neil CAN SLIM
Macro / liquidity	Regime, policy, dollar liquidity, risk appetite	Druckenmiller, Soros, PTJ, Dalio, Arthur Hayes, 李蓓, 时寒冰
Value / quality	Moat, owner earnings, margin of safety	Buffett, Munger, Li Lu, Nick Sleep, 段永平, 邱国鹭, 张磊, 但斌
Deep value / special sits	Forced sellers, spinoffs, distress	Graham net-net, Schloss, Klarman, Greenblatt, Tepper, Templeton
Growth / innovation	Duration, S-curves, unit economics	Lynch, Cathie Wood, Coleman / Tiger, Sequoia, a16z, YC
Activist / forensic	Governance, accounting, shorts	Icahn, Ackman, Einhorn, Muddy Waters, Hindenburg, Spruce Point
Crypto structure	Narrative, on-chain, funding, invalidation	Ansem, Hsaka, CryptoCred, Willy Woo, Cobie, Paradigm, Multicoin, Delphi, Placeholder
A-share / China systems	Cycle, board strength, volume-price	冯柳, 邓晓峰, 唐能通, 陈浩
Quant / systematic	Many weak edges, cost, decay	Simons

Open-source investor skills (catalog)

These packages are designed to run inside Questflow’s harness (and in Codex / Claude / Cursor):

Western classic: Livermore, O’Neil CAN SLIM, Buffett, Munger, Howard Marks, Soros, Druckenmiller, Darvas, Minervini VCP, Burry, Simons, Lynch, Klarman, Greenblatt, Tepper, PTJ, Seykota, Turtles, Icahn, Li Lu, Nick Sleep, Dalio, Gundlach, Ackman, Einhorn, Coleman, Cathie Wood, Graham, Schloss, Templeton, Hohn, Watsa.

Crypto: Ansem, Serenity, Arthur Hayes, Hsaka, CryptoCred, Willy Woo, Cobie, Paradigm, Multicoin, a16z crypto, Delphi, Placeholder.

Venture lenses: Sequoia, a16z, USV, Founders Fund, YC.

China: 段永平, 冯柳, 邱国鹭, 邓晓峰, 但斌, 李蓓, 唐能通, 陈浩, 张磊, 时寒冰.

Forensic / short: Muddy Waters, Hindenburg, Spruce Point, Einhorn.

How skills get used

1. A fund manager’s judgment is distilled into an AI Agent (often composing several skills).

2. The agent reads plugins for data.

3. A model reasons inside the skill’s rules.

4. Execution happens only inside your venue permissions and risk limits.

Skills do not invent alpha. They structure judgment so it can be monitored and executed.

Related: Four pillars · What is invest.md? · What is trading DNA?

Nothing here is financial advice. A skill is a framework, not a signal service.