Critical Vulnerabilities in Open-Source Software Rise Nearly Fivefold
From January to September 2026, CodeScoring recorded nearly five times as many critical open-source vulnerabilities as a year earlier. Site owners and teams should check the versions of the components they use.
On November 6, 2026, CodeScoring published the results of a study of vulnerabilities in open-source software. According to CodeScoring data published on its Habr blog, the platform’s database added records of 3,566 critical vulnerabilities from January to September—4.9 times as many as the 723 recorded over the same period in 2025. These are issues with a CVSS score of 9.0 or higher. Site owners and teams should review what their applications contain and which component versions they use, especially in live, internet-facing services.
How the npm and Go Figures Changed
In npm, the package registry for JavaScript, the number of recorded critical vulnerabilities rose from 89 to 562—a 6.3-fold increase. The number of stable package versions affected by these issues grew even faster, from 3,844 to 39,127, or 10.2 times as many.
For Go, researchers recorded 396 critical vulnerabilities, compared with 78 over the same period last year. That is a 5.1-fold increase. The number of affected stable module versions reached 12,443, up from 5,995—a 2.1-fold increase. Both ecosystems therefore saw growth in the number of known issues and the number of versions they affect, though at different rates.
For more than two-thirds of the critical vulnerabilities recorded in 2026, the CVSS assessments indicate that an attack can be carried out over a network without prior access privileges or user interaction. In this scenario, an attacker does not need to persuade an employee to open a file or hand over credentials: the target is the network-accessible software itself. This does not mean that every issue listed is present on any given website, but it explains the focus on internet-facing services.
The database continued to grow outside the nine-month period selected for the study. From September 15 to October 15, 2026, it added records of 4,609 new, unique vulnerabilities in open-source components. This figure covers all severity levels, not just issues with a CVSS score of 9.0 or higher. It cannot be compared directly with the January–September total: both the time period and the scope differ.
Why More Vulnerabilities Are Known
CodeScoring attributes the rise in published vulnerability records primarily to more active defect discovery and broader tracking. Researchers use automated analysis and AI-based tools, while developers disclose information about discovered issues more often. The figures therefore show an increase in known vulnerabilities. On their own, they do not prove that 4.9 times as many new bugs appeared in open-source code over the year.
Improved testing methods also help uncover issues in code that has been in use for years. A vulnerability may be reported years after a library was incorporated into a product and the team stopped updating it. The absence of obvious problems in an application does not mean new advisories will not appear for its components.
The study also does not claim that all the vulnerabilities listed are already being exploited or pose the same risk to every website. The figures describe published records concerning open-source components, not the state of a specific project. To determine whether a report affects a live product, a team needs to compare it with what is actually installed in the application.
What Site Owners and Teams Should Check
The first step is to identify which open-source components—and exactly which versions—are used in the applications a website or service depends on. A package name alone is not enough: vulnerability reports identify affected versions. If the software inventory is unknown, it is hard to tell whether a new report calls for action. A one-off inventory will not solve the problem for long, either: new information continues to emerge even for libraries adopted years ago.
CodeScoring recommends software composition analysis. It can automatically track what applications contain, identify vulnerable dependencies, and determine which component versions fix an issue. This helps teams separate reports relevant to their product from the wider stream of vulnerability disclosures. Checks should cover not only the software as a whole, but also the components it contains.
Issues affecting live, internet-facing services should take priority. For a site owner, this is a matter of keeping the product running; for a developer, it helps identify which dependency to check and update first. A marketer does not need to understand every CVSS score, but it is useful to know whether the team can regularly check the dependencies of the platform the business uses to communicate with customers.
My View: Dependencies Need Ongoing Attention
I agree that a package list is not a minor technical detail to revisit only before launch. I would start with a simple question about my project: can I quickly tell whether a new vulnerability affects a live service? If answering it means reconstructing the application’s dependencies by hand, it is time to change the review process.
At the same time, I would not treat the “nearly fivefold” figure as a reason to update everything indiscriminately. To me, it is a signal to establish regular checks and deal first with issues in software accessible from outside. That turns vulnerability reports into specific tasks for the team rather than a stream of alarming notifications.
Sources
Where the news comes from. The text is a retelling in the author’s own words; the facts come from the source, the opinion is the author’s.
full-stack developer
I build web products for small teams, from APIs to interfaces, mostly with TypeScript. In my spare time, I try new approaches on a personal project—right now it’s a map of quiet places to work in Cyprus.
All posts by the authorRelated articles
Malicious Script Hits Laravel Package and Thousands of GitHub Repositories
A malicious script reached a Laravel package and GitHub Actions workflows. Traces of the campaign were found in 8,926 repositories. Teams should review build history and the secrets available to their workflows.
543,000 Active Secrets Found in Public GitHub Repositories
Truffle Security analyzed 224 million public GitHub repositories and found 543,000 active secrets. Many had been exposed for years, while some go undetected by GitHub’s built-in push protection.
Audit of 284 Russian Websites Finds Email, DNS and Data Protection Issues
An audit of 284 websites run by major Russian organizations found issues with DMARC, DNS, outdated libraries and third-party services. Automated findings still need manual verification.
Discussion 4
Levon Vardanyan
One thing worth adding: a version check is only the first pass. In a small Node app, I’d also check whether the affected package is actually reachable from production code, then update and run the tests; otherwise a big count of vulnerable package versions can make the risk look more immediate than it is. Does CodeScoring distinguish reachable issues from packages that are merely present in the dependency tree?
Vasilis Iacovou Post author
good distinction — the figures in the post count recorded vulnerabilities and affected versions, not whether a package is reachable in a specific app. I’d check reachability in the production dependency tree, then patch and run tests; that gives a much better signal than the version count alone.
Ewa Nowicka
In a small learning project, I once updated a package just because the audit output looked scary, then spent longer fixing a breaking change than checking whether that code path was used at all. Now I start with the production dependency tree, patch what’s relevant, and run the tests — your note about reachability is a useful reminder that the version counts are a starting point, not the whole risk picture.
Elise Evers
Do you know if CodeScoring’s figures include a reachability check, or are they counts of vulnerable versions in the dependency data? That distinction matters before treating the 39,127 affected npm versions as production risk.