The invitation landed in my DMs like clockwork. A recruiter from a top-tier Layer-2 protocol, offering a remote role with a generous token allocation. The catch? A mandatory AI-driven interview process using a new tool called "Relay." I almost clicked. Two days later, SlowMist published their analysis: Relay is a custom infostealer targeting Web3 professionals, and dozens have already been compromised.
This is not a new chapter in the same old phishing story. It is a weaponized convergence of social engineering and cross-platform malware, designed to drain not just wallets but the very trust that binds our ecosystem. And it works because we, the idealists of Web3, have forgotten that trust—real trust—must be earned through technical rigor, not blind community spirit.
Context: The Perfect Phishing Environment
The crypto job market has always operated on relationships. In 2020, during DeFi Summer, I joined a Discord server, found a lead, and interviewed over a voice call. No formal HR, no background check. That culture of informality was beautiful—but it also created a treasure map for attackers. Today, with remote work normalized and AI recruitment tools flooding the market, the attack surface has exploded. Scammers clone legitimate job postings, scrape LinkedIn profiles of real recruiters, and craft tailored invitations that pass even a skeptical eye.
SlowMist's report details the attack chain: a fake recruiter reaches out, often citing a specific project you've contributed to or written about. They push you to install "Relay"—a piece of software that claims to use AI to synchronize meeting notes and schedule interviews. The bait is irresistible: cutting-edge tech for a cutting-edge job. Once installed, the malware executes a silent info-stealing routine, targeting: - Browser credentials (including saved passwords for exchanges, wallets, and email) - Crypto wallet files (MetaMask, Phantom, Exodus, and others) - macOS Keychain (storing seeds, private keys, and API tokens) - Telegram session tokens (giving attackers full access to your chats and groups) - System information for lateral movement
This is not a script kiddie operation. The malware has been compiled for both macOS and Windows, with code obfuscation and anti-analysis techniques. The attackers have invested in cross-platform compatibility—a clear signal they expect a high-value, tech-savvy target set.
Core Insight: What This Attack Reveals About Web3
1. The Human Layer Is Our Weakest Protocol
We obsess over zero-knowledge proofs, sharding, and optimistic rollups. But the most elegant smart contract cannot protect a user who voluntarily hands over their private key to a piece of malware. This attack exploits a cognitive vulnerability: the optimism bias that attracts people to Web3. We believe in a better financial system, so we trust the people who claim to help build it.
I remember 2017, drowning in whitepapers, installing random wallet extensions to test every new ICO. That same eagerness now opens the door. The attack preys on the very emotions that make us crypto natives: curiosity, ambition, and a desire to be first.
2. Cross-Platform Malware Means No Safe Haven
Mac users have long felt immune. "I can't get a virus," they say. Relay shatters that myth. It uses macOS-specific APIs to access the Keychain and browser storage, and it persists via LaunchAgents. On Windows, it uses more traditional techniques like scheduled tasks and registry modifications. The message is clear: no operating system is safe if you run unverified executables.
3. Telegram Sessions: The Backdoor to Your Community
The theft of Telegram session tokens is particularly insidious. Telegram is the nerve center of most Web3 projects—for governance discussions, admin chats, and private deal rooms. Once an attacker gains access, they can impersonate the victim, request funds from colleagues, or inject malicious links into trusted groups. This turns one compromised individual into a supply chain attack on the entire project.
I have seen it happen. In 2022, a friend lost control of their Telegram after a similar session cookie theft. The attacker then convinced four community members to send ETH for a "private sale." The losses multiplied. This is not just an individual risk; it is a systemic one.
4. The Privacy Paradox Meets AI
As we integrate more AI tools into our workflow, we trade convenience for exposure. The same language models that help us write code can be used to craft more convincing phishing messages. The same tools that automate our interviews can be weaponized to steal our keys. This mirrors a deeper tension in Web3: our desire for sovereignty versus our need for usability.
Resilience is the new utility. We cannot halt innovation, but we must layer security into every new tool. That means demanding transparency from software providers, verifying code signatures, and maintaining a strict separation between work and financial machines.
Contrarian Angle: Why Hardware Wallets Are Not Enough
The immediate advice from security experts: "Use a hardware wallet." True, but incomplete. If your computer is compromised, a hardware wallet can protect against remote private key extraction, but it cannot prevent an attacker from signing a transaction you did not intend. When you receive a pop-up on an infected machine, the attacker can replace the recipient address with their own. Your hardware wallet will display the fake address, and you will approve it.
The real solution is more radical: never run unverified software on the same machine you use for significant crypto operations. Create a dedicated environment for interviews—a virtual machine, a live USB, or a cheap laptop that holds no financial data. Treat every new application as a potential attack vector until proven otherwise.
This is a culture shift. In the bear market, we learned to hoard and hodl. In this new phase, we must learn to isolate and verify. Trust is built in the bear, sold in the bull—but it can be stolen in a single careless install.
Takeaway: From Ashes to Armor
SlowMist's disclosure is a gift. It gives us the IOCs to detect this specific malware and the chance to harden our habits before the next variant emerges. But the broader lesson is about our collective mindset. Web3 was born from a desire to trust code over humans. Yet we have built a human-centric culture that thrives on collaboration and openness. That culture must now evolve, not by becoming cynical, but by becoming technically disciplined.
From the ashes of 2022, we planted seeds for 2030. Those seeds need armor—a layer of skepticism, a commitment to isolated environments, and a community-wide standard for verifying identity. The promise of decentralization can only survive if we protect the people behind the keys.