← Back to Skills Marketplace
543
Downloads
1
Stars
0
Active Installs
9
Versions
Install in OpenClaw
/install vibe-trading
Description
Professional finance research toolkit — backtesting (7 engines + benchmark comparison panel), factor analysis, options pricing, 72 finance skills, 29 multi-a...
Usage Guidance
This skill appears to do what it says, but it requires you to pip-install and run a third-party MCP server (vibe-trading-ai / vibe-trading-mcp). Before installing:
- Verify the PyPI package and its author (view the project on PyPI/GitHub, inspect source code) — pip installs can run arbitrary code on your machine.
- Prefer installing inside an isolated environment (virtualenv, container, or disposable VM) and test there first.
- Review what the MCP server does: read its docs/source to confirm network endpoints it opens, data it logs, and whether it exfiltrates data.
- Only provide optional API keys (TUSHARE, OPENAI, broker credentials) if you trust the package and limit their permissions where possible.
- If you need stronger guarantees, run the backtests and file-processing steps locally with manual commands instead of registering the MCP server, or request a vetted binary/release signed by a known maintainer.
If you can't inspect the package source or verify the maintainer, treat this skill as higher-risk and avoid installing it on sensitive machines.
Capability Analysis
Type: OpenClaw Skill
Name: vibe-trading
Version: 0.1.6
The vibe-trading bundle provides a comprehensive finance toolkit but includes high-risk capabilities such as arbitrary file system access (read_file, write_file) and network communication (web_search, read_url). It specifically targets sensitive financial data by requesting users to provide broker CSV exports for its 'Shadow Account' analysis. While these features are consistent with the stated goal of financial research, the combination of broad system access and the processing of private financial records via the vibe-trading-ai dependency presents a significant security risk without further inspection of the underlying package code.
Capability Tags
Capability Assessment
Purpose & Capability
Name/description (finance research, backtesting, multi-agent teams) align with the declared tools and listed features in SKILL.md. Optional env vars (TUSHARE_TOKEN, OPENAI_API_KEY, LANGCHAIN_MODEL_NAME) map to described optional features (China A-shares, run_swarm). No unrelated credentials or binaries are requested.
Instruction Scope
SKILL.md instructs the user/agent to pip install the package, add an MCP server entry that runs the package's 'vibe-trading-mcp' command, and to use tools that read user-provided CSV trade exports and create files (write_file). Those actions are coherent for the stated functionality but grant the installed package and its MCP process broad local capabilities (file I/O, network calls, running servers). The instructions do not explicitly request unrelated system files or secrets, but the agent will be told to run a third-party process that can access those resources.
Install Mechanism
There is no platform install spec, but the SKILL.md tells users to run 'pip install vibe-trading-ai' from PyPI. Installing an unreviewed PyPI package and running its MCP server spawns arbitrary code on the host. This is expected for CLI/MCP-style skills but is a higher-risk install pattern than instruction-only skills that don't require new software.
Credentials
Declared env vars are optional and directly tied to features (TUSHARE for China A-share data; OPENAI API and model name for multi-agent swarms). The skill does mention optional broker login (Futu) for some data; these are reasonable and proportional to the functionality described. No unrelated secrets are requested.
Persistence & Privilege
The skill is not always-enabled and defaults allow autonomous invocation (normal). The notable persistence aspect is the MCP server: adding 'vibe-trading-mcp' to MCP config causes the agent runtime to spawn and trust a locally-installed process that can handle requests. That increases blast radius compared with an instruction-only skill but is consistent with MCP-based plugin design.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install vibe-trading - After installation, invoke the skill by name or use
/vibe-trading - Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.1.6
v0.1.6: Fixes #55 (swarm presets bundled inside src.swarm package — vibe-trading --swarm-presets now works after pip/uv tool install). Plus AKShare ETF/forex routing fix. Rolls up benchmark panel + upload safety + Futu loader + vnpy export skill since v0.1.5.
v0.1.5
Shadow Account: extract broker rules → backtest shadow → P&L leakage report (4 new tools + 1 skill). Trade Journal Analyzer: 4 behavior bias diagnostics (disposition / overtrading / momentum-chasing / anchoring). Universal File Reader: pdf/docx/xlsx/pptx/images/40+ text formats in one envelope. Agent Harness v2: persistent memory + FTS5 session search + self-evolving skills. 22 MCP tools, 71 skills, 29 swarm presets.
v0.1.4
Add web_search MCP tool, akshare/ccxt data sources, fix Docker build, bump to 68 skills
v0.2.3
Remove bundled mcp_server.py — MCP runs via pip-installed vibe-trading-mcp command
v0.2.2
Clarify API key requirements: 15/16 MCP tools work with zero keys, add per-tool key column
v0.2.1
Clarify API key requirements: 15/16 MCP tools work with zero keys, add per-tool key column to tools table
v0.2.0
Add vibe-trading-mcp entry point, document package vs commands
v0.1.1
Add get_market_data, get_swarm_status, get_run_result, list_runs tools
v0.1.0
Initial public release of vibe-trading: a professional finance research toolkit.
- Adds 64 specialized finance skills covering backtesting, factor analysis, risk, options, and more.
- Includes 29 multi-agent swarm teams for investment research and quant strategy.
- Supports data and backtesting for HK/US equities (via yfinance), cryptocurrency (via OKX), and China A-shares (via Tushare; key optional).
- Provides 16 MCP tools for workflows such as skill discovery, multi-agent execution, data fetch, file handling, and report generation.
- No API keys required for HK/US/crypto; optional TUSHARE_TOKEN for China A-shares.
Metadata
Frequently Asked Questions
What is Vibe Trading?
Professional finance research toolkit — backtesting (7 engines + benchmark comparison panel), factor analysis, options pricing, 72 finance skills, 29 multi-a... It is an AI Agent Skill for Claude Code / OpenClaw, with 543 downloads so far.
How do I install Vibe Trading?
Run "/install vibe-trading" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Vibe Trading free?
Yes, Vibe Trading is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Vibe Trading support?
Vibe Trading is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Vibe Trading?
It is built and maintained by Haozhe Wu (@warren618); the current version is v0.1.6.
More Skills