Fake Trading Libraries: 228 Malicious PyPI and npm Packages Posing as Exchange and Trading Tools

Last updated: 20 October 2026 Are fake trading libraries a real risk? We searched the OpenSSF malicious-packages database, published through OSV, for packages whose names copy exchanges, brokers or trading tools, then checked each name by hand to remove false matches.

Key findings · OSV malicious-package records to 20 Oct 2026

228 malicious packages pretending to be trading or exchange tools. After checking every name by hand, we count 183 on npm (JavaScript) and 45 on PyPI (Python) in the OpenSSF malicious-packages database, reported from 2022 to 20 October 2026.

Polymarket is the new favourite target. 50 packages use Polymarket or prediction-market names, 48 of them reported in 2026, many posing as copies of Polymarket’s official CLOB trading client or as trading-bot “Kelly stake” calculators.

2025 and 2026 are the busiest years. 92 were reported in 2025 and 76 so far in 2026. Coinbase (35) and Binance (32) names are the next most copied.

The usual goal is your keys. Reports describe packages that run code on install and send off environment variables, .env files and API keys, or download remote malware. A trading bot’s exchange or broker API keys are exactly what they look for.

Malicious trading packages by year

Year reportedPyPInpmTotal
202201515
202327633
20244812
202558792
202696776

Years are when the OSV record was published, which is usually within days of the package appearing. Most of these packages were removed by npm or PyPI soon after being reported, so the risk is mainly to anyone who installed them during that window.

Which names they copy

Name targetsPackages
Polymarket / prediction markets50
Coinbase35
Binance32
Generic trading tools19
Huobi / HTX13
Kraken10
MEXC8
KuCoin8
Bitget8
Hyperliquid6
CCXT (trading library)6
OKX5
Bybit5
Bitstamp4
Bitfinex4
Freqtrade (trading bot)4
TradingView3
BitMEX3
Deribit2
yfinance (market data)2
Robinhood1

Recent examples (non-Polymarket, since June 2026)

ReportedRegistryPackageOSV record
2026-09-22PyPIcrypto-trader-pyMAL-2026-16406
2026-09-09npmgmgn-trading-kitMAL-2026-16094
2026-09-05npmtrading-bot-utilsMAL-2026-16003
2026-08-31npmhyperliquid-composerMAL-2026-15627
2026-08-14npm@mexc/shared-utilsMAL-2026-14049
2026-08-14npm@devmikets/hyperliquid-sdkMAL-2026-14042
2026-08-05npmgekko-trading-botMAL-2026-12388
2026-08-05npmccxt-tradingMAL-2026-12347
2026-08-05PyPIcrypto-trading-toolkitMAL-2026-12082
2026-07-13npmgamified-trading-systemMAL-2026-10502
2026-07-09npm@wagni_bot/binance-sdkMAL-2026-10022
2026-07-09npm@wagni_bot/hyperliquid-sdkMAL-2026-10026
2026-07-09npm@wagni_bot/hyperliquidMAL-2026-11111
2026-06-29npm@huobi-ui/activity-componentsMAL-2026-6627
2026-06-10npmcoinbase-wallet-utilsMAL-2026-5499

All records are in the full CSV with links. A package name here does not mean the real company, exchange or library was involved; these are lookalikes published by unrelated accounts.

How the attacks work

  • Typosquats. Names one letter off from real libraries, such as pyhton-binance for python-binance or cccxt for ccxt.
  • Lookalike SDKs. Unscoped copies of official packages, such as polymarket-clob-client instead of @polymarket/clob-client.
  • Dependency confusion. Names that look like a company’s internal packages, such as coinbase-internal or kraken-production, hoping a build system pulls the public one.
  • Fake tools. Packages sold as trading bots, copy-trading tools or AI trading agents that run hidden code on install.

How to protect your trading setup

  • Install exchange and broker SDKs only under the exact name in the official documentation, and check the publisher.
  • Keep API keys out of the folder where you install packages, and give keys trade-only permissions with no withdrawals.
  • Pin versions with a lock file and review new dependencies before installing.
  • If you installed any package on this list, revoke your API keys and wallet keys first, then clean the machine.

For the legitimate libraries traders use and how popular they are, see our Python algo trading package downloads and JavaScript trading library npm downloads pages.

Download the dataCSV: all 228 packages with registry, OSV ID, date reported, which name they copy, who reported it, and a link.
Download CSV ↓

Sources

Method

We took every MAL- record in the OSV PyPI and npm exports and matched package names against exchange, broker and trading terms (for example binance, coinbase, kraken, polymarket, hyperliquid, ccxt, freqtrade, tradingview, trader, trading).

That gave 359 matches. We reviewed each by hand and removed 128 that only matched by accident, such as random strings containing “okx” or “mt4”, a spam campaign of names containing “alpaca” the animal, and “kraken” names unrelated to the exchange. We also set aside 3 packages labelled as security-research proofs of concept. 228 remain.

Caveats

OSV only includes packages someone reported, so the real number is higher. Name matching misses malicious packages with unrelated names that still target traders.

Records are reports by security firms and researchers listed in the CSV. We haven’t analysed the code ourselves.

How to cite this

How to cite · CC BY 4.0

Prop Data Lab, “Malicious Trading Packages on PyPI and npm (OSV)”, updated 20 October 2026, https://propdatalab.com/malicious-trading-packages-pypi-npm/

The table and the CSV are free to reuse under CC BY 4.0: copy, chart or republish them, commercial use included, as long as you link back to this page. Data: OpenSSF Malicious Packages via OSV (Apache-2.0).

Common questions

Are there malicious Python trading libraries?

Yes. We found 45 malicious PyPI packages with names copying trading tools or exchanges, such as typosquats of python-binance and ccxt, reported to the OSV database from 2023 to October 2026.

Is there a fake Polymarket npm package?

The OSV database lists 50 malicious packages with Polymarket or prediction-market names, including lookalikes of Polymarket's official CLOB client. The official package is published under the @polymarket scope.

What do malicious trading packages steal?

Reports describe packages that collect environment variables, .env files and API keys on install, or download and run remote malware.

How do I check if a package is malicious?

Search the package name on osv.dev, confirm the publisher matches the official documentation, and avoid new packages with few downloads that copy a well-known name.