← Back to list
mean-reversion Checked
Mean-reversion strategy tools including Hurst exponent, half-life estimation, z-score signals, AD…
8.8Overall rating
Installation
🤖 Install via AI
Copy the prompt below and send it to your AI assistant (e.g. Claude Code) — it will follow the instructions and install automatically.
Install the "mean-reversion" skill by following the instructions at https://skill123.me/install/mean-reversion.
⌨️ Command line install
Run in your terminal — downloads and installs to ~/.claude/skills/.
curl -fsSL https://skill123.me/install/mean-reversion.sh | bash
📦 Download ZIP
Download the zip and extract it into your skills directory (e.g. ~/.claude/skills/), then restart your session.
⬇ Download v1.0.0 · 24 KBAbout this skill
Overview
Mean-reversion strategy tools including Hurst exponent, half-life estimation, z-score signals, ADF testing, and Ornstein-Uhlenbeck modeling
Source
- Repo: https://github.com/agiprolabs/claude-trading-skills
- Path: skills/mean-reversion
Score breakdown
Trigger
7.0
Description enumerates the statistical toolkit with strong keywords (Hurst exponent, half-life, z-score, ADF, Ornstein-Uhlenbeck); When/When-Not absent from description but well covered in body (ranging vs trending markets, regime changes, liquidity traps) — partial credit.
Structure
8.7
304-line SKILL.md with two well-scoped references and two scripts; inline stubs correctly defer full implementations to references; Files section annotates contents.
Workflow
8.8
Excellent workflow discipline: statistical confirmation gate before trading, complementary three-test battery, half-life-derived lookback/holding/stop rules, stop at abs(z)>3 (reversion failed), regime filtering via sibling skill, and demo-mode verification in both scripts.
Content
9.0
Formula-forward, table-driven, minimal filler; every parameter has a rule of thumb.
Engineering
10.0
Frontmatter valid; shebangs, dependencies, and env vars documented; all referenced paths exist.
Security
10.0
6.0 manifest: sole network destination is public-api.birdeye.so (official, declared, optional live mode); BIRDEYE_API_KEY via env only; no file writes, no credential reads, no persistence; explicit not-financial-advice disclaimer present. Capabilities match declared scope.
Statistically rigorous mean-reversion skill: ADF/Hurst/variance-ratio test battery with a test-before-trade gate, half-life estimation with rule-of-thumb parameter derivation, z-score signal framework with stops, OU parameter estimation, and honest crypto caveats (Hurst 0.55-0.70 for majors). Body includes explicit When-It-Works/When-It-Fails sections and a financial-advice disclaimer. Scripts default to synthetic demo mode. Deductions: description lacks Use-when phrasing; some SKILL.md code stubs defer to references (correct disclosure but slightly fragmented).