AGENT SKILL
Data & Analytics74 trading skills for process and risk, explicitly not for signals
tradermonty/claude-trading-skills
2,867 stars, MIT, Japanese README included. 74 skills across 8 categories, 58 production and 16 beta, all governed by one canonical skills-index.yaml where each entry carries a verification matrix (data source, required or recommended, standalone or workflow_step, maturity); the README catalogue is generated from it. YAML manifests under workflows/ chain the skills into executable flows enforced by a validator.
Editor's takeWhat earns trust is the self-limitation: it states it is not designed for fully automated trading, signal outsourcing or scalping, a long disclaimer tops the README, and the licence is MIT 'AS IS'. The swing workflow runs drawdown-circuit-breaker and pre-trade-discipline-gate before any screener — risk gates before tickers. That ordering is the soul of the design and the opposite of most tools that lead with a list of picks.
git clone https://github.com/tradermonty/claude-trading-skills && cp -r claude-trading-skills/skills/backtest-expert ~/.claude/skills/Positioning: process and risk, not stock signals
tradermonty/claude-trading-skills (2,867 stars / 651 forks, MIT, docs at tradermonty.github.io, Japanese README included) is a trading research toolkit for time-constrained individual investors: a long-term core (ETF/dividend) plus a disciplined satellite swing book. It states repeatedly that it is not financial advice, not a signal service, not an order-execution platform — it governs process, risk and journalling. A long disclaimer tops the README; MIT, "AS IS, WITHOUT WARRANTY".
Saying who it is NOT for matters just as much: it is not designed for fully automated trading, signal outsourcing, or short-term scalping. That self-limitation is rare in trading tools and is exactly what makes it credible.
74 skills, 8 categories, 58 production + 16 beta
Every skill is governed by one canonical metadata index, skills-index.yaml (id, category, integrations, workflow back-references); the README's catalogue table is auto-generated by generate_catalog_from_index.py. Scale and distribution:
| Category | Count | Representative skills |
|---|---|---|
| market-regime | 16 | market-breadth-analyzer, uptrend-analyzer, exposure-coach, ibd-distribution-day-monitor, macro-regime-detector |
| strategy-research | 15 | backtest-expert, manifoldbt-backtester, edge-pipeline-orchestrator |
| meta | 9 | Navigation and meta capabilities (Navigator, quality dashboard) |
| swing-opportunity | 7 | vcp-screener, canslim-screener, breakout-trade-planner |
| trade-planning | 7 | position-sizer, technical-analyst, pre-trade-discipline-gate |
| core-portfolio | 7 | portfolio-manager, value-dividend-screener, kanchi-dividend-sop |
| advanced-satellite | 7 | parabolic-short-trade-planner, earnings-trade-analyzer, options-strategy-advisor |
| trade-memory | 6 | trader-memory-core, signal-postmortem |
Each skill carries a verification matrix in the index (which data source, required vs recommended, standalone vs workflow_step, production vs beta). Structuring dependency and maturity that way keeps 74 skills from turning into a pile.
workflows/: strings skills into executable routines
What makes it usable is the YAML manifests under workflows/ — each one names the exact skills, decision gates and artifacts in order, enforced by a validator (--strict-workflows). New users start from these five:
| Goal | Workflow | Anchor skills | API profile |
|---|---|---|---|
| 15-minute daily market check | market-regime-daily | market-breadth-analyzer, uptrend-analyzer, exposure-coach | No API for the basic path |
| Weekly long-term portfolio review | core-portfolio-weekly | portfolio-manager, kanchi-dividend-review-monitor, trader-memory-core | Alpaca required; manual CSV is a degraded fallback |
| Find swing candidates only when risk allows | swing-opportunity-daily | vcp-screener, drawdown-circuit-breaker, technical-analyst, position-sizer, trader-memory-core, pre-trade-discipline-gate | FMP for screeners; local state for the risk and discipline gates |
| Record and learn from every closed trade | trade-memory-loop | trader-memory-core, signal-postmortem | No API for the manual path |
| Monthly performance review and rule adjustment | monthly-performance-review | trader-memory-core, signal-postmortem, backtest-expert | No API for the manual path |
Note the two gates in the swing workflow: drawdown-circuit-breaker and pre-trade-discipline-gate. Clear risk first, then look for candidates — that ordering is the soul of the design and the opposite of most tools that hand you a list of tickers first.
Install and companion
# Claude web app: download the .skill from skill-packages/, upload via Customize → Skills
# Claude Code: clone the repo, copy the skill folders you want into ~/.claude/skills/ (or project-level .claude/skills/)
git clone https://github.com/tradermonty/claude-trading-skills
cp -r claude-trading-skills/skills/backtest-expert ~/.claude/skills/The repo also has skillsets/ (goal-based bundles of required/recommended/optional skills; four shipped first: market-regime, core-portfolio, swing-opportunity, trade-memory, consumed by the Navigator) and a companion Hermes Trading Research Agent Work Package that wraps these skills into slash commands like /pre-market-routine, /after-close-review, /trade-journal, /weekly-portfolio-review and /monthly-performance-review. It repeats: this is a research, journalling and risk-review assistant that does not place orders, run a signal service, or run hidden scheduled jobs — human decision gates stay central.
Platform support is Python 3.10–3.13 (Linux/Windows/macOS), with a 3.9 source floor for packaged standalone skills; only combinations exercised by CI count as "supported".