F6 Family Position Reverse
This skill describes the F6 commodity futures factor research workflow. It uses Panda data broker net margin, dominant contract and daily futures data to compute a cross-sectional reverse signal from selected family-style brokers.
Runtime Entry
- Codex: read
AGENTS.md, then thisSKILL.md. - Claude Code: read
CLAUDE.md, then thisSKILL.md. - Cursor: load
.cursor/rules/skill-alpha-f6-family-position-reverse.mdc. - Hermes: read
HERMES.md, then thisSKILL.md. - OpenClaw: read
OPENCLAW.md, then thisSKILL.md.
Skill Contents
- Development skill:
alpha-f6-family-position-reverse/SKILL.md - Production-read skill:
alpha-f6-family-position-reverse-production/SKILL.md - Data guide:
alpha-f6-family-position-reverse/references/data_guide.md - Offline smoke test:
alpha-f6-family-position-reverse/test.py
Required Commands
Run from alpha-f6-family-position-reverse/:
python test.py
python scripts/factor.py
python scripts/validate.py
python scripts/backtest.py
python test.py uses synthetic data and does not require Panda data
credentials. The other commands require PANDA_DATA_USERNAME and
PANDA_DATA_PASSWORD.
Boundaries
This repository is research and tooling material. It is not investment advice, does not promise returns, and is not an official endorsement by QuantSkills, Panda data, Codex, Claude Code, Cursor, Hermes or OpenClaw.
The recommended public repository name is
skill-alpha-f6-family-position-reverse.