Guide · Agent skill
How to use the Smart Money Concepts Signals agent skill (smc)
Updated · written with AI from the project's own documentation and checked against our catalogue data
The smc skill teaches an AI agent a specific chart-analysis method called Smart Money Concepts (also called ICT, from the Inner Circle Trader school). It looks for structure breaks, fair value gaps and order blocks, then returns a simple signal of 1 (long), -1 (short) or 0 (stand aside). This guide explains what the skill does, how to install it, how to ask for it, and what it needs and risks, so you can use it for study and prototyping.
Smart Money Concepts SignalsHKUDS/Vibe-Trading · MIT Download the skill Quick start
- 1Download the skill ZIP. Get the smc ZIP from this page and unzip it.
- 2Place the folder. Put the smc folder in ~/.claude/skills/ for personal use, or in a project's .claude/skills/ folder.
- 3Or upload in the Claude apps. Go to Settings → Capabilities → Skills and upload the skill.
- 4Install the Python dependencies. Run pip install "vibe-trading-ai[smc]" along with pandas, numpy and requests.
- 5Ask in plain language. Request an SMC analysis by naming BOS, ChoCH, FVG or order blocks and the data you want.
- 6Review the output. Check that the result is 1, -1 or 0 and read the reasoning before relying on it.
What is an agent skill, and what does the smc skill make the agent do?
An agent skill is a folder of instructions and optional scripts that an AI agent can load when a task matches the skill's description. The smc skill is a Smart Money Concepts (ICT) signal engine. When it is active, the agent applies this method to price data instead of improvising its own approach.
The method is built on a few concepts: BOS (break of structure) as a trend continuation signal, ChoCH (change of character) as a trend reversal signal, FVG (fair value gap) as a price refill target, and order blocks as zones where institutional orders are thought to concentrate. Liquidity is also described as a stop-hunt target zone.
How does the smc skill decide a signal?
The skill sets direction from ChoCH, confirms it with BOS, and filters it with FVG. A long needs a bullish ChoCH or BOS plus a bullish FVG. A short needs a bearish ChoCH or BOS plus a bearish FVG. If there is no structural signal, or the directions conflict, the answer is to stand aside.
The reference notes say ChoCH carries the most weight because it signals a change in market structure, BOS is used for confirmation, and the FVG filter is meant to reduce false signals.
- 1 = long
- -1 = short
- 0 = stand aside
- Parameters: swing_length (default 10) sets the window for swing high and low detection; close_break (default True) decides whether BOS/ChoCH requires a closing-price break.
How do you install the smc skill?
Download the ZIP for the skill from this page and unzip it. For a personal setup, place the smc folder in ~/.claude/skills/. For a single project, place it in that project's .claude/skills/ folder. Alternatively, in the Claude apps, go to Settings → Capabilities → Skills and upload the skill there.
The skill also needs the Python library. The source lists this install command for the dependencies:
- Python with pandas, numpy and requests
- The smartmoneyconcepts Python library
pip install "vibe-trading-ai[smc]"How do you trigger the skill with natural requests?
You do not need a special command. Describe the analysis you want in plain language, and mention the concepts by name so the agent matches the skill. The skill's description covers Smart Money Concepts, ICT, BOS, ChoCH, FVG and order blocks.
Good requests name the market, the data and the output you want. For example, you might ask the agent to check an OKX BTC candle series for a ChoCH with an FVG filter and report 1, -1 or 0. Keep requests to research and testing.
- "Use the Smart Money Concepts skill to scan this price data for BOS and ChoCH and tell me whether the signal is 1, -1 or 0."
- "Explain why the FVG filter rejects this long signal."
- "Run the smc example engine on OKX candles and show the signal convention used."
What does the skill include and what does it need?
The skill folder includes SKILL.md (the instructions), example_signal_engine.py (an example engine that can download OKX crypto candle data and compute signals), and two reference notes in Chinese on structure breaks and on order blocks and fair value gaps. The source marks the skill as running scripts and using the network.
For the example data fetch, the skill needs Internet access to www.okx.com. The source states that no API key appears in the files.
- Python with pandas, numpy and requests
- smartmoneyconcepts Python library, installed via the pip command above
- Internet access to www.okx.com for the example data fetch
- Files: example_signal_engine.py, references/结构突破.md, references/订单块与缺口.md, SKILL.md
What are the safety notes?
The skill executes Python and makes outbound web requests. Run it only on data you trust, and do not give it live exchange credentials. Review example_signal_engine.py before running it so you know what it will fetch and execute.
The signals are a rule-based sketch. The skill gives no risk management, no position sizing and no evidence of profitability. It should not be used to make real trading decisions with money.
- Run scripts only in an environment you control.
- Do not supply live exchange credentials to the agent.
- Treat outputs as study material, not investment advice.
What tips help you get useful results?
Check the parameters before you trust a result. Changing swing_length changes which swing highs and lows the engine detects, and close_break changes whether a wick alone can count as a break. Read the two reference notes to see how the agent reasons about each concept.
Ask the agent to explain each signal in terms of ChoCH, BOS and FVG. That makes it easier to spot where a result comes from and whether it matches the method.
- Start with the default parameters, then test one change at a time.
- Ask for the reasoning behind every 1, -1 or 0.
- Compare results across several time periods before drawing conclusions.
What people use it for
- Studying how ICT-style market structure rules are coded step by step.
- Prototyping a backtest that turns SMC signals into 1, -1 or 0 values.
- Comparing BOS and ChoCH signals across OKX crypto candle data for research.
- Teaching students how institutional-style concepts like FVG and order blocks translate into code.
Frequently asked questions
Is the smc skill free to use?
The source does not state a price or licence terms beyond the item's listed licence, which is MIT. The skill depends on the smartmoneyconcepts library and Python packages, which you install yourself. Check the licence of each dependency before using it commercially.
What do I need to run the smc skill?
You need Python with pandas, numpy and requests, the smartmoneyconcepts library installed with pip install "vibe-trading-ai[smc]", and internet access to www.okx.com for the example data fetch. The source says no API key appears in the files.
Where do I install the smc skill?
Unzip it into ~/.claude/skills/ for personal use or into a project's .claude/skills/ folder. You can also upload it in the Claude apps under Settings → Capabilities → Skills.
Does the smc skill tell me when to buy or sell?
It produces a rule-based signal of 1, -1 or 0. The source says it gives no risk management, position sizing or performance validation, and it is not evidence of profitability, so it should not drive real trades.
Why does the skill return 0?
Zero means stand aside. The skill returns it when there is no structural signal from ChoCH or BOS, or when the directions conflict, and it also needs an FVG that matches the direction for a long or short.
What should I use instead of smc if I want a tested trading system?
The source does not name alternatives. For a validated system, you would need your own backtesting, risk controls and performance validation, which this skill explicitly does not provide.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use the smc skill if you already study ICT-style price action, want to learn how SMC concepts are coded, or need a starting point for a backtest or research prototype. Pick something else if you want investment advice, risk management or proof of profitability, since the source provides none of these and the skill runs scripts and network calls.
Download the skill