What does this skill do, and when should you use it?
This is an MCP server that brings live market data (stocks, crypto, forex, futures), 30+ technical indicators, global screeners, 9-strategy backtesting with walk-forward overfitting detection, Reddit sentiment, and financial news to AI assistants like Claude, ChatGPT, and Cursor. It can be self-hosted via pip or used through a hosted version. The skill configures an AI agent to call these tools directly via a bash wrapper, enabling natural-language trading queries across multiple markets.
The skill executes a Python wrapper (trading.py) that calls the tradingview-mcp server's 37 tools. It fetches real-time quotes (yahoo_price), bulk prices for multiple symbols (get_prices_bulk), and global market snapshots (market_snapshot). It runs technical analyses (RSI, MACD, Bollinger Bands, Supertrend, etc.), backtests 9 strategies with institutional metrics and optional trade logs/equity curves, performs walk-forward validation for overfitting detection, analyzes Reddit sentiment, fetches financial news, and scans markets for bullish/oversold stocks. It supports US, crypto, Turkish BIST, and Egyptian EGX markets. Outputs are formatted for concise display, suitable for Telegram.
- A trader asks for real-time price, change, and 52-week range for AAPL and gets a compact summary.
- An investor compares the performance of RSI vs. Supertrend strategies on BTC-USD over one year to decide which to use.
- An analyst runs a walk-forward backtest on an RSI strategy to check if it is overfitted, receiving a ROBUST/OVERFITTED verdict.
- A crypto enthusiast queries Reddit sentiment on BTC, getting a bullish/bearish score and top posts.
- A day trader scans NASDAQ for strong bullish or oversold stocks to build a watchlist.
- A user connects the skill to Telegram via OpenClaw and sends commands like 'market snapshot' or 'backtest AAPL RSI 1 year'.
How do you install this skill?
- The skill depends on overseas data sources (Yahoo Finance, TradingView, Reddit) which may be unreachable from mainland China; evaluate network accessibility.
- The skill executes commands directly without user confirmation, potentially performing actions without explicit consent; use caution when handling financial data.
- Static review cannot verify actual functionality; test the skill yourself and be aware of data accuracy and latency.
- The skill does not disclose the complete data flow; understand how data is processed and transmitted.
- Shell / CLI
- Network access
- Local filesystem
- MCP Server
Python 3.10-3.13uvtradingview-mcp-server package
Self-hosted: 1) Install Python 3.10-3.13 and uv (curl -LsSf https://astral.sh/uv/install.sh | sh); 2) Run 'pip install tradingview-mcp-server' or clone repo and 'uv run tradingview-mcp'; 3) Optionally set MARKETAUX_API_TOKEN for news/sentiment; 4) Place the skill folder in ~/.agents/skills/tradingview-mcp/ and ensure SKILL.md is at openclaw/SKILL.md. Hosted: Sign up at pro.cryptosieve.com, get a connector URL, and paste it into Claude.ai or ChatGPT.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/atilaahmettaner/tradingview-mcp.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/openclaw" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- Backtest RSI strategy on BTC-USD for 1 year
- Compare all strategies on MSFT for 2 years
Ensure the skill is installed and the trading.py wrapper is at ~/.openclaw/tools/trading.py. In an AI chat, ask trading questions directly, e.g., 'What is AAPL's price?', 'Backtest RSI strategy on BTC-USD for 1 year', or 'Compare all strategies on MSFT for 2 years'. The AI agent will run commands via bash: python3 ~/.openclaw/tools/trading.py <command>. Note: SKILL.md explicitly says never use sessions_spawn or ask for an agent ID.
What are this skill's strengths and limitations?
- Covers diverse asset classes (stocks, crypto, forex, ETFs, indices) across global exchanges
- Includes 9 backtest strategies with institutional metrics (Sharpe, Calmar, Profit Factor) and walk-forward overfitting detection
- No API keys required (except optional Marketaux for news), free and open source (MIT)
- Multi-platform support (Claude, ChatGPT, Cursor, Copilot) and OpenClaw (Telegram, WhatsApp, etc.)
- Combines technical analysis with Reddit sentiment and live news for confluence decisions
- Hosted version available for zero-setup access
- Self-hosting requires Python and uv setup; first-run can hit dependency or network issues (e.g., Python 3.14 unsupported)
- Market data may be delayed or inaccurate; not financial advice
- Backtest results depend on data source; no official test suite, testing coverage unknown
- Output for Telegram is concise, losing detailed JSON information
- Hosted version costs $9-$29/month; free self-hosting requires manual maintenance and updates
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| TradingView MCP AI Trading Intelligence this page | 50 · Use with care | ★ 5k | 4d ago | MIT |
| OpenNews Financial News Aggregator | 47 · Use with care | ★ 2.4k | 1mo ago | MIT |
| Market Intel — AI-Trader Read-Only Market Data Skill | 47 · Use with care | ★ 23k | 4mo ago | — |
| Pattern Recognition Bot (2.1) | 42 · Not recommended | ★ 2.1k | 1mo ago | AGPL-3.0 |
| Global Stock Data — Full-Stack US & HK Market Data Skill | 62 · Recommended | ★ 1.7k | 4d ago | Apache-2.0 |
Compared to Bloomberg Terminal, this tool is free or low-cost and sets up in minutes, offering similar data and analysis capabilities. Unlike DIY scripting for technical analysis, this integrates many features, saving development time.
How did FollowSkills review this skill?
Evidence: Skill invokes a local script via bash with no external transmission, privilege escalation, or sensitive data access; no user confirmation; depends on external data sources (Yahoo Finance, TradingView) but no credentials; MIT license with clear attribution. Deductions: lack of user confirmation; incomplete disclosure of data-flow paths; insufficient isolation (all commands execute directly in the same environment); rollback not addressed.
Evidence: trading.py provides argument validation (_require_symbol) and error capture, outputting JSON error messages; SKILL.md details command usage and defaults; README includes troubleshooting. Deductions: static review cannot execute; dependence on live external services means no determinism; only catches exceptions without retries or degradation; error messages may not be fully clear to users.
Evidence: Skill clearly targets market data analysis with listed symbols (stocks, crypto, ETFs, indices, Turkish, Egyptian); trigger conditions are clear (user asks for price, technical analysis, etc.); language detection supports Chinese; but core function relies on overseas services (Yahoo Finance, TradingView, Reddit) which may be unreachable in mainland China. Deductions: capability boundaries not fully defined (e.g., data delays, market limitations); environmental fit in mainland China is constrained without alternatives.
Evidence: SKILL.md is well-structured with quick-reference tables and example formats; README provides detailed installation, configuration, troubleshooting; MIT license, version 0.8.0, clear maintenance responsibility (Atila Ahmettaner). Deductions: SKILL.md lacks installation/dependency notes (it depends on external trading.py); no known-limitation disclosure (e.g., data source reliability); version history only partially in README, not in SKILL.md.
Evidence: Skill claims to provide real-time prices, technical analysis, backtesting, sentiment, and has clear output formats (Telegram-style) with examples. Deductions: static review cannot verify actual outputs; no independent test evidence (e.g., unit test suite covering key paths); limited evidence of comparative benefit (relies on external services; hosted version is paid, self-host free).
Evidence: Test file (tests/stress) exists but for stress testing, not key-path unit tests; pyproject configures pytest but no CI running evidence; README demo GIFs/screenshots are not independently verifiable. Deductions: no third-party execution evidence, no CI workflow file (e.g., GitHub Actions) showing test runs; key claims (e.g., 37 tools, 9 strategies) lack independent verification.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →