Oracle3 prediction-market agent
MCP server (13 tools) and 6 agent skills for Kalshi and Polymarket: search markets, read rules and order books, compare related contracts, and compute fee-adjusted net edge, with a local paper ledger. Live execution is a separate Python engine.
- Scan
- Repo scan pass
- GitHub stars
- 475
- Last updated
- 2026-10-12
- License
- Apache-2.0
- Cost
- free
Why it helps
Lets a forecasting agent research and paper-test cross-market mispricings with fees counted, without exchange credentials or real orders until the owner opts in.
Ask your owner first. installs a mcp
- Keep it on paper mode until the paper-trading-discipline gates are met.
Paste to your agent
Read https://upskill.coach/mcp/oracle3-prediction-markets.md. If the security scan passed, ask me before installing, then install it with the command under Install and tell me what you did.Install
git clone the repo, pip install -e . in a venv, then add .venv/bin/oracle3-mcp as an MCP server (e.g. claude mcp add oracle3 -- <path>/oracle3-mcp).
Community
Helps with
Repo security-scanned on 2026-10-19 at commit 267fb7c48dae: pass. The verdict covers the whole repo (yichengyang-ethan/oracle3-prediction-market-agent); this mcp wasn't scanned separately.
- Scan notes
- Automated scan: pass (medium 1) at commit 267fb7c48dae on 2026-10-19.
- Source
- github.com/yichengyang-ethan/oracle3-prediction-market-agent
- Repo
- yichengyang-ethan/oracle3-prediction-market-agent
Static scan at a pinned commit (prompt-injection, malware and exfiltration rules plus semgrep, gitleaks, osv-scanner); flagged lines reviewed by hand. A scan lowers risk; it doesn't prove safety.
Alternatives
Compare allApify MCP Server (official)
Hosted (mcp.apify.com) or local MCP server exposing Apify Actors: scrapers for Google Maps, LinkedIn, X, Instagram, websites, e-commerce.
vectorbt
A fast vectorized backtesting engine for testing many trading strategy variants on historical data.
backtesting.py
A small Python framework to backtest trading strategies, with stats and interactive plots.
Firecrawl MCP Server (official)
MCP for scraping, crawling, search and structured extraction (hosted v2 endpoint or local).
Related
prediction-market-analysis
A framework for collecting and analyzing prediction-market data; its README lists research papers that used it.
Prediction Market Agent Tooling (Gnosis)
Gnosis Python library to benchmark, deploy and monitor prediction-market agents: AbstractBenchmarkedAgent vs crowd wisdom on Manifold/etc., DeployableTraderA…
Forecast calibration scoring
Log every probability forecast, then score resolved ones with the Brier score and log loss, against the market, with a calibration table.
vectorbt
A fast vectorized backtesting engine for testing many trading strategy variants on historical data.
backtesting.py
A small Python framework to backtest trading strategies, with stats and interactive plots.
Financial Datasets MCP server
An MCP server for the Financial Datasets stock-market API: financial statements, prices and news.
Comments
Comments are from visitors, not reviewed by Upskill, and never shown to agents.