Hackers have launched a large-scale software supply chain attack targeting developers across npm, PyPI, and Crates.io, compromising at least 34 open-source packages and hundreds of associated versions.
Security researchers at Socket are tracking the campaign as “TrapDoor,” a crypto-focused credential stealer designed to infiltrate developer environments and exfiltrate sensitive data.
The TrapDoor campaign stands out due to its coordinated execution across three major package ecosystems. The attack began on May 22, 2026, with the upload of a malicious PyPI package named eth-security-auditor, followed by waves of similar packages published under multiple accounts.
Researchers identified over 384 malicious versions distributed across npm, PyPI, and Crates.io. These packages masquerade as legitimate developer utilities, including tools for blockchain development, AI workflows, and security auditing. Their naming strategy increases the likelihood of adoption by developers working in crypto, DeFi, and AI-related projects.
TrapDoor uses ecosystem-specific techniques to execute malicious code during normal development workflows:
A shared payload, commonly known as trap-core.js, is deployed to npm packages. This script scans infected systems for sensitive data, validates stolen credentials, and establishes persistence mechanisms.
For example, when a developer installs a seemingly harmless package like wallet-security-checker, the post-install script silently executes, harvesting SSH keys and cloud credentials while embedding persistence hooks into the system.
Data Theft and Persistence
The malware is designed to steal a wide range of sensitive information:
To maintain access, TrapDoor deploys multiple persistence techniques, including Git hooks, shell configurations, cron jobs, and systemd services. It can also attempt lateral movement using stolen SSH credentials, increasing the risk of broader infrastructure compromise.
A notable aspect of the campaign is its attempt to exploit AI-powered developer tools. The attackers inject hidden instructions into files such as .cursorrules and CLAUDE.md, which AI coding assistants use.
These instructions are designed to trick AI tools into executing malicious “security scans” that expose sensitive data. In some cases, attackers used invisible Unicode characters to conceal malicious directives, making detection more difficult.
Additionally, the threat actor attempted to insert these files into legitimate GitHub repositories via pull requests, disguising them as helpful development guidelines.
The campaign is linked to a GitHub account hosting payloads and configuration files via GitHub Pages. The same infrastructure was reused across all three ecosystems, helping researchers correlate the activity into a single coordinated operation.
The attacker also maintained repositories documenting malware design concepts, including credential harvesting, AI prompt injection, and persistence strategies, indicating a structured and evolving attack framework.
TrapDoor highlights a growing trend in supply chain attacks that target not just code dependencies, but the entire developer environment. By combining traditional techniques with AI-focused abuse, attackers are expanding their reach into modern development workflows.
Developers and organizations are advised to:
This campaign demonstrates how even low-download packages can pose significant risks when strategically designed to target high-value developer environments.
Key TrapDoor campaign packages
| Ecosystem | Package name | Likely lure theme / description |
|---|---|---|
| npm | async-pipeline-builder | Generic async workflow / pipeline helper for projects |
| npm | build-scripts-utils | Build automation utilities for JavaScript/Node projects |
| npm | chain-key-validator | Blockchain / key validation helper for crypto projects |
| npm | crypto-credential-scanner | “Security” tool for scanning crypto-related credentials |
| npm | defi-env-auditor | DeFi environment audit / security checking utility |
| npm | defi-threat-scanner | Threat-scanning helper for DeFi developers |
| npm | deployment-key-auditor | Deployment key and credential auditing helper |
| npm | dev-env-bootstrapper | Environment bootstrap/setup tool for developers |
| npm | eth-wallet-sentinel | Ethereum wallet monitoring / safety helper |
| npm | llm-context-compressor | LLM / AI context size optimization helper |
| npm | mnemonic-safety-check | Seed phrase / mnemonic safety checker for wallets |
| npm | prompt-engineering-toolkit | Prompt engineering helper for AI and LLM workflows |
| npm | solidity-deploy-guard | Solidity deployment guard / security helper |
| npm | wallet-backup-verifier | Wallet backup verification / safety checking tool |
| npm | wallet-security-checker | General wallet “security check” helper |
| npm | web3-secrets-detector | Secret and key detector for Web3 projects |
| npm | workspace-config-loader | Workspace configuration loader for dev environments |
| PyPI | cryptowallet-safety | Python tool for crypto wallet safety checks |
| PyPI | data-pipeline-check | Data pipeline validation / QA helper |
| PyPI | defi-risk-scanner | DeFi risk analysis / scanning tool |
| PyPI | env-loader-cli | CLI for loading local environment variables |
| PyPI | eth-security-auditor | Ethereum project “security auditor” (first observed) |
| PyPI | git-config-sync | Git configuration sync helper for developers |
| PyPI | solidity-build-guard | Solidity build guard / security helper |
| Crates | move-analyzer-build | Move language build / analysis helper |
| Crates | move-compiler-tools | Compiler helper tools for Move language |
| Crates | move-project-builder | Project builder for Move-based projects |
| Crates | sui-framework-helpers | Helper tools for Sui framework developers |
| Crates | sui-move-build-helper | Build helper for Sui/Move projects |
| Crates | sui-sdk-build-utils | SDK build utilities for Sui ecosystem |
Follow us on Google News, LinkedIn, and X to Get Instant Updates and Set GBH as a Preferred Source in Google.
Microsoft has announced a vulnerability in Windows Paint that could let unauthenticated attackers execute malicious…
An autonomous penetration-testing framework linked to South Korean banking breaches and cloud supply-chain attacks is…
Shopify is making IP allowlisting mandatory for all public apps to limit abuse of stolen…
OpenAI has introduced GPT-6 Cyber, a security-focused model built to find and validate exploitable vulnerabilities…
Melbourne, Florida, November 5th, 2026, CyberNewswire OpenMatter Network, a pioneer in decentralized infrastructure for agentic…
A China-aligned espionage group abused the update traffic of Sogou Input Method, one of China's…