Trivy Supply Chain Attack Triggers Self-Spreading CanisterWorm Across 47 npm Packages

Trivy Supply Chain Attack Triggers Self-Spreading CanisterWorm Across 47 npm Packages Ravie LakshmananMar 21, 2026Malware / Threat Intelligence The threat actors behind the supply chain attack targeting the popular Trivy scanner are suspected to be conducting follow-on attacks that have led to the compromise of a large number of npm packages with a previously undocumented self-propagating worm dubbed CanisterWorm. The name is a reference to the fact that the malware uses an ICP canister, which refers to tamperproof smart contracts on the Internet Computer blockchain, as a dead drop resolver. The development marks the first publicly documented abuse of an ICP canister for the explicit purpose of fetching the command-and-control (C2) server, Aikido Security researcher Charlie Eriksen said. The list of affected packages is below - 28 packages in the @EmilGroup scope 16 packages in the @opengov scope @teale.io/eslint-config @airtm/uuid-base32 @pypestream/floating-ui-dom The development comes within a day after threat actors leveraged a compromised credential to publish malicious trivy, trivy-action, and setup-trivy releases containing a credential stealer. A cloud-focused cybercriminal operation known as TeamPCP is suspected to be behind the attacks. The infection chain involving the npm packages involves leveraging a postinstall hook to execute a loader, which then drops a Python backdoor that's responsible for contacting the ICP canister dead drop to retrieve a URL pointing to the next-stage payload. The fact that the dead drop infrastructure is decentralized makes it resilient and resistant to takedown efforts. "The canister controller can swap the URL at any time, pushing new binaries to all infected hosts without touching the implant," Eriksen said. Persistence is established by means of a systemd user service, which is configured to automatically start the Python backdoor after a 5-second delay if it gets terminated for some reason by using the "Restart=always" directive. The systemd service masquerades as PostgreSQL tooling ("pgmon") in an attempt to fly under the radar. The backdoor, as mentioned before, phones the ICP canister with a spoofed browser User-Agent every 50 minutes to fetch the URL in plaintext. The URL is subsequently parsed to fetch and run the executable. "If the URL contains youtube[.]com, the script skips it," Eriksen explained. "This is the canister's dormant state. The attacker arms the implant by pointing the canister at a real binary, and disarms it by switching back to a YouTube link. If the attacker updates the canister to point to a new URL, every infected machine picks up the new binary on its next poll. The old binary keeps running in the background since the script never kills previous processes." It's worth noting that a similar youtube[.]com-based kill switch has also been flagged by Wiz in connection with the trojanized Trivy binary (version 0.69.4), which also reaches out to the same ICP canister via a Python dropper ("sysmon.py"). As of writing, the URL returned by the C2 is a rickroll YouTube video. The Hacker News found that the ICP canister supports three methods – get_latest_link, http_request, update_link – allowing the threat actor to modify the behavior at any time to serve an actual payload. In tandem, the packages come with a "deploy.js" file that the attacker runs manually to spread the malicious payload to every package a stolen npm token provides access to in a programmatic fashion. The worm, assessed to be vibe-coded using an artificial intelligence (AI) tool, makes no attempt to conceal its functionality. "This isn't triggered by npm install," Aikido said. "It's a standalone tool the attacker runs with stolen tokens to maximize blast radius." To make matters worse, a subsequent iteration of CanisterWorm detected in "@teale.io/eslint-config" versions 1.8.11 and 1.8.12 has been found to self-propagate on its own without the need for manual intervention. Unlike "deploy.js," which was a self-contained script the attacker had to execute with the pilfered npm tokens to push a malicious version of the npm packages to the registry, the new variant incorporates this functionality in "index.js" within a findNpmTokens() function that's run during the postinstall phase to collect npm authentication tokens from the victim's machine. The main difference here is that the postinstall script, after installing the persistent backdoor, attempts to locate every npm token from the developer's environment and spawns the worm right away with those tokens by launching "deploy.js" as a fully detached background process. Interestingly, the threat actor is said to have swapped out the ICP backdoor payload for a dummy test string ("hello123"), likely to ensure that the entire attack chain is working as intended before adding the malware. "This is the point where the attack goes from 'compromised account publishes malware' to 'malware compromises more accounts and publishes itself,'" Eriksen said. "Every developer or CI pipeline that installs this package and has an npm token accessible becomes an unwitting propagation vector. Their packages get infected, their downstream users install those, and if any of them have tokens, the cycle repeats." (This is a developing story. Please check back for more details.) Found this article interesting? Follow us on Google News, Twitter and LinkedIn to read more exclusive content we post. SHARE     Tweet Share Share SHARE  Blockchain, cybersecurity, DevOps Security, Malware, NPM, Open Source Security, Python, supply chain attack, Threat Intelligence Trending News ThreatsDay Bulletin: OAuth Trap, EDR Killer, Signal Phishing, Zombie ZIP, AI Platform Hack and More Meta to Shut Down Instagram End-to-End Encrypted Chat Support Starting May 2026 OpenClaw AI Agent Flaws Could Enable Prompt Injection and Data Exfiltration Microsoft Patches 84 Flaws in March Patch Tuesday, Including Two Public Zero-Days Critical n8n Flaws Allow Remote Code Execution and Exposure of Stored Credentials Six Android Malware Families Target Pix Payments, Banking Apps, and Crypto Wallets Chinese Hackers Target Southeast Asian Militaries with AppleChris and MemFun Malware FortiGate Devices Exploited to Breach Networks and Steal Service Account Credentials Google Fixes Two Chrome Zero-Days Exploited in the Wild Affecting Skia and V8 ⚡ Weekly Recap: Chrome 0-Days, Router Botnets, AWS Breach, Rogue AI Agents and More Researchers Trick Perplexity's Comet AI Browser Into Phishing Scam in Under Four Minutes CISA Flags Actively Exploited Wing FTP Vulnerability Leaking Server Paths Nine CrackArmor Flaws in Linux AppArmor Enable Root Escalation, Bypass Container Isolation Apple Issues Security Updates for Older iOS Devices Targeted by Coruna WebKit Exploit Android 17 Blocks Non-Accessibility Apps from Accessibility API to Prevent Malware Abuse Veeam Patches 7 Critical Backup and Replication Flaws Allowing Remote Code Execution Apple Fixes WebKit Vulnerability Enabling Same-Origin Policy Bypass on iOS and macOS Load More ▼ Popular Resources Fix Security Noise by Focusing Only on Validated Exposures Get the 2026 ASV Report to Benchmark Top Validation Tools Guide - Discover How to Validate AI Risks With Adversarial Testing Webinar - Identify Key Attack Paths to Your Crown Jewels with CSMA