{"slug": "supply-chain-zero-days-closing-the-gap-between-publication-and-detection", "title": "Supply Chain Zero Days: Closing the Gap Between Publication and Detection", "summary": "Wiz built a supply chain detection pipeline that analyzes new npm package versions at publication, using static analysis, an AI classifier and sandboxed dynamic analysis to flag malicious packages and add detection rules in Wiz Defend before a CVE is published. Wiz Research reported that notable supply chain attacks rose from about 10% of significant incidents in H2 2025 to 25% in H1 2026, and cited the Axios compromise, in which an unknown threat actor published two malicious npm versions (v1.14.1 and v0.30.4) depending on the newly created plain-crypto-js package. Axios was present in roughly 80% of cloud and code environments and downloaded about 100 million times per week, with observed execution in 3% of affected environments.", "body_md": "Supply Chain Zero Days: Closing the Gap Between Publication and Detection\n\nWhen trusted packages are compromised, malicious code can quickly spread throughout and beyond developer environments. See how Wiz’s supply chain pipeline identifies compromised packages at publication, helping teams detect zero-day supply chain compromises in Wiz Defend before a CVE is published\n\nModern software development depends on open source packages. Developers and AI coding agents can pull new dependencies and versions into projects at speed, helping teams build and ship faster.\n\nBut that speed creates a challenge for security teams. When a trusted package is compromised, malicious code can enter development environments and build pipelines before traditional vulnerability databases flag the threat.\n\nOrganizations cannot manually review or block every update to foundational packages without slowing development. Therefore, most organizations scan newly downloaded packages against known vulnerabilities and block if needed. But, zero-day supply chain compromises won’t show up in vulnerability databases right away; and, because popular packages are already trusted, developer workflows and build pipelines are often configured to automatically pull their latest versions. If an attacker compromises one of those packages, a malicious update can spread across the environment before a CVE is even published, leaving security teams to defend from behind.\n\nTrusted package compromises in the wild\n\nOur own research team is seeing this technique in practice more often than ever before. In H2 2025, notable supply chain attacks made up about 10% of significant incidents. In H1 2026, that percentage increased to 25%, and now, notable supply chain attacks account for 1 in 4 significant security incidents. During the Axios compromise, for example, an unknown threat actor compromised the npm account of an axios maintainer and published two malicious versions of the npm package (v1.14.1, v0.30.4), which introduced a dependency on plain-crypto-js, a newly created malicious package.\n\nAlthough the malicious versions were removed within hours, that didn't mean that the threat disappeared. According to the Wiz Research team’s own observations, at the time, Axios was present in approximately 80% of cloud and code environments, and downloaded around 100 million times per week. The package’s popularity enabled rapid exposure, with observed execution in 3% of affected environments. This reality underlines the importance of speed in defending against supply chain attacks.\n\nDetecting malicious packages at publication\n\nTraditionally, security teams find out a trusted package was compromised by receiving a threat advisory or vulnerability database update. In an ideal world, teams could take a proactive approach to securing the software supply chain by knowing which packages published to npm are clean, preventing installation of those that are not, and alerting on malicious versions across all tools.\n\nThat’s why we built a dedicated supply chain detection pipeline to analyze new packages and versions as soon as they are published to npm. This allows us to flag malicious packages and add detection rules in Wiz Defend, code and cloud pipelines, all before a formal CVE is published.\n\nHere’s how it works:\n\nWhen a new package version is released, the pipeline downloads and extracts the current version and previous versions for comparison.\n\nStatic analysis techniques flag suspicious files. This analysis feeds into our AI classifier engine, which returns an initial verdict.\n\nSuspicious versions undergo dynamic analysis to validate the verdict. This dynamic analysis involves in depth work to understand the file contents, including:\n\nInstalling the package version in a sandbox environment to monitor runtime activity.\n\nImporting the version and exposed modules .\n\nRunning the package from the CLI (where possible).\n\nThe Wiz Sensor observes runtime behavior to monitor for malicious activity. This helps the pipeline confirm the suspicious verdict, and provides our AI classifier with additional evidence.\n\nThe AI classifier returns the final verdict, generated based on all previous analyses. Once a version is confirmed as malicious, the pipeline automatically publishes IOCs to the Wiz reputation service. This ensures that Wiz customers are alerted as soon as a package enters their environment, via:\n\nThreat intelligence center notifications, alerting on compromised packages so teams can understand if they are impacted.\n\nWiz Defend, with auto-updated detection rules for malicious packages and configurable runtime detection and response policies.\n\nCombining various analysis methods and live runtime execution allows us to flag zero-day compromises at speed while keeping false positives to a minimum.\n\nDetection in the real world\n\nA notable example of the pipeline in action is the keyv attack, where multiple npm packages in the keyv/cacheable ecosystem were compromised following the compromise of a GitHub maintainer account. In this particular case, instead of hardcoding command-and-control (C2) domains into the package source code, where static rules could easily extract and block them, the malware dynamically retrieved its C2 domains at runtime by querying an Ethereum smart contract. Because the malicious domains never appeared in the code itself, traditional static analysis alone could not identify the C2 infrastructure.\n\nUsing the supply chain pipeline, the Wiz Research Team was able to successfully see network activity to the malicious domain, automatically detecting the malicious domain connection without human intervention. Combining static and dynamic analysis, along with AI classification and live runtime monitoring, helped the pipeline flag the package with high confidence. As soon as the package was confirmed malicious, the relevant information automatically pushed to Wiz Defend and across the Wiz platform, enabling customers to detect compromised keyv versions in their environments before a formal CVE was published.\n\nProactive detection of compromised packages with Wiz\n\nOpen source packages are fundamental to the modern software development ecosystem. As workflows accelerate, security teams need real-time visibility and control over third-party dependencies before malicious code can land on developer workstations or build pipelines.\n\nBy analyzing new npm package versions as they are published and distributing confirmed threat intelligence across the platform, Wiz helps teams identify compromised packages before traditional vulnerability signals become available. In addition, Wiz evaluates dependencies during each install or upgrade process, to help security teams detect and manage risky or compromised package versions at the exact point of ingestion.\n\nThe goal is simple: empower security and development teams to move at machine speed, with the visibility and control over what enters the software supply chain.", "url": "https://wpnews.pro/news/supply-chain-zero-days-closing-the-gap-between-publication-and-detection", "canonical_source": "https://www.wiz.io/blog/detect-supply-chain-zero-days-at-publication", "published_at": "2026-10-19 14:39:29+00:00", "updated_at": "2026-10-19 15:15:39.182421+00:00", "lang": "en", "topics": ["ai-safety", "ai-tools", "ai-infrastructure"], "entities": ["Wiz", "Wiz Research", "Wiz Defend", "Axios", "npm", "plain-crypto-js", "Wiz Sensor"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/supply-chain-zero-days-closing-the-gap-between-publication-and-detection", "markdown": "https://wpnews.pro/news/supply-chain-zero-days-closing-the-gap-between-publication-and-detection.md", "text": "https://wpnews.pro/news/supply-chain-zero-days-closing-the-gap-between-publication-and-detection.txt", "jsonld": "https://wpnews.pro/news/supply-chain-zero-days-closing-the-gap-between-publication-and-detection.jsonld"}}