Candlestick Pattern Recognition
Scans price candles for 15 classic chart patterns and produces long/short/neutral signals
- Repository
- HKUDS/Vibe-Trading
- Stars
- 37,433
- Licence
- MIT
- Files
- 2 files
What it does
This skill teaches an AI agent to read price charts the way many traders do, by spotting 15 well-known candlestick shapes (such as hammers, engulfing bars, and morning/evening stars) in price data. It scores bullish patterns +1 and bearish ones -1, then turns the total into a simple signal: long, short, or stand aside. The included Python file shows how this can be computed quickly across a whole price history using pandas.
Great for
- Quantitative or technical-analysis learners who want to see how candlestick rules are coded
- Developers building a trading-strategy prototype who need a starting pattern library
- Analysts who want a quick, rules-based scan of historical OHLC price data
Skip it if
- Making real trading decisions with money: the pattern rules are simple, unvalidated by the files, and ignore costs, risk management and market context
- Investment advice for beginners who may trust the signal without backtesting, since the skill gives no evidence of profitability
Try asking
- “Scan my BTC daily OHLC data and tell me which bullish or bearish candlestick patterns appeared in the last 30 days”
- “Write a pandas function that flags engulfing patterns in a CSV of stock prices”
- “Explain what a morning star candlestick pattern is and how the signal score works”
- “Combine candlestick scores with a moving-average filter for a simple backtest”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 2 original files, the MIT licence and an attribution note.
candlestick.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip candlestick.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip candlestick.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Validate any signal on historical data with a backtest before relying on it; the skill itself provides no performance results.
- Adjust body_pct and shadow_ratio carefully, since default thresholds strongly affect which candles count as patterns.
- Ask the agent to explain each detected pattern with its date so you can check it visually against a chart.
What's inside
- example_signal_engine.py
- SKILL.md
Redistributed unmodified from HKUDS/Vibe-Trading under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.