FindSpot risk in code and cloud
ProveConfirm what's exploitable
FixShip fixes that get merged
Nullify never merges. A developer approves every fix.
Learn
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Suspendisse varius enim in eros elementum tristique. Duis cursus, mi quis viverra ornare, eros dolor interdum nulla, ut commodo diam libero vitae erat. Aenean faucibus nibh et justo cursus id rutrum lorem imperdiet. Nunc ut sem vitae risus tristique posuere.
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Suspendisse varius enim in eros elementum tristique. Duis cursus, mi quis viverra ornare, eros dolor interdum nulla, ut commodo diam libero vitae erat. Aenean faucibus nibh et justo cursus id rutrum lorem imperdiet. Nunc ut sem vitae risus tristique posuere.
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Suspendisse varius enim in eros elementum tristique. Duis cursus, mi quis viverra ornare, eros dolor interdum nulla, ut commodo diam libero vitae erat. Aenean faucibus nibh et justo cursus id rutrum lorem imperdiet. Nunc ut sem vitae risus tristique posuere.
Exploited CVEs, supply-chain compromises, leaked secrets, app-layer breaches and many more.
A heap overflow in HotSpot's hardware-accelerated Base64 decoder can be triggered without authentication, but only on HotSpot Java 17.0.6+, 21, 25 or 27 running on x86-64 CPUs with AVX-512 VBMI.
Read the analysis →An attacker exploited a vulnerability in Optimizely Experimentation to add malicious JavaScript to customer projects on November 16, 2026, serving a crypto clipper, card skimmer and iOS exploit kits from 178 organizations' sites.
Read the analysis →Coles disclosed a cyber security incident to the ASX on November 16, 2026. It says it has found no evidence that customer, team or supplier data was compromised, and it hasn't said what happened.
Read the analysis →Hijacked GitHub accounts pushed a secret stealer into thousands of repositories, and an npm hook got past GitHub's workflow hold.
Read the analysis →An authentication bypass chained with an OS command injection lets attackers take over Barracuda Email Security Gateway appliances, and CISA gave federal agencies three days to patch.
Read the analysis →An attacker exploited a remote code execution flaw in SGLang, the open-source inference engine, on one NEAR AI server and moved into wider infrastructure.
Read the analysis →GPUStack trusted requests that looked like they came from localhost as admin, so a reverse proxy or a forged header gave attackers full control.
Read the analysis →A compromised Kernel API key led an attacker into a logging provider, where it downloaded some customer API keys and deployment secrets, then onto GPU servers for cryptomining.
Read the analysis →Twelve releases across ten hijacked npm packages carried a runtime loader that finds its command server through Ethereum transactions and can change its payload remotely.
Read the analysis →One SSO flaw led to forged staff sessions, an over-scoped GitHub App credential and production secrets.
Read the analysis →An attacker obtained root API keys for some Trigger.dev environments, which exposed their environment variables, affecting 0.66% of customers.
Read the analysis →An attacker targeted secrets held in Northflank's control plane, and Northflank told every customer to assume their environment variables were decrypted and rotate all keys.
Read the analysis →A crafted regular expression in a GitLab CI configuration could trigger an integer overflow in Ruby's regex engine and run code on self-managed servers.
Read the analysis →A scan of 7.25 million public Gitee repositories found 26,400 unique credentials that were still live, led by cloud and AI provider keys.
Read the analysis →An unrestricted file upload flaw in Adobe Commerce and Magento Open Source can lead to code execution in a user's context, and CISA lists it as exploited.
Read the analysis →A tool tracked as Gheist used stolen GitHub tokens to plant sync-check workflows that printed CI secrets into logs and deleted the evidence within seconds.
Read the analysis →A compromised release of the popular LLM gateway shipped a credential stealer to PyPI and LiteLLM's container registries.
Read the analysis →Two critical bugs in the PHP interpreter allowed unauthenticated remote code execution, and the vulnerable version lives in container images and servers, not in composer.lock.
Read the analysis →An attacker escalated privileges in Railway's build infrastructure, pulled encrypted environment variables and decrypted a subset, affecting up to about 100,000 workspaces.
Read the analysis →A path traversal flaw across Atlassian Data Center products lets unauthenticated attackers read specific files in the web application root.
Read the analysis →An unauthorized party used internal Fireworks AI credentials to read environment variables used in evaluation jobs, likely capturing customer API keys stored as secrets.
Read the analysis →tensorlake 0.5.144 on npm shipped a Shai-Hulud-style credential stealer that wipes the home directory if the stolen GitHub token is revoked too early.
Read the analysis →Hijacked GitHub accounts pushed secret-stealing workflows into tens of thousands of repositories, and later variants mined full git history for credentials.
Read the analysis →At least nine AI developer platforms disclosed incidents in October 2026, and attackers hunted the environment variables that hold AI provider keys.
Read the analysis →A modified release workflow swapped @subql/common for an attacker tarball, so version 5.8.3 shipped with credential theft and a remote shell and valid provenance.
Read the analysis →An unauthenticated path traversal in WordPress core's page-template resolution can include local PHP files and lead to code execution on some servers.
Read the analysis →Republished MemTensor packages on npm and PyPI launched a bundled binary called sckit from runtime code paths, which gets around disabled install scripts.
Read the analysis →A path traversal and missing authentication in GitLab's repository commits API let unauthenticated users read files from self-managed servers.
Read the analysis →A template injection flaw in Adobe Commerce and Magento Open Source allows unauthenticated code execution, and Adobe says it's being exploited in the wild.
Read the analysis →An outsider abused a comment-triggered release workflow to publish ten malicious versions of @7nohe/openapi-react-query-codegen through npm Trusted Publishing.
Read the analysis →Malicious versions of arrayref, internment and append-only-vec added a build-time dependency that downloaded and ran a payload during cargo build.
Read the analysis →An unauthenticated SSRF in MLflow's webhook test endpoint lets attackers reach internal services and cloud metadata and read the responses.
Read the analysis →A flaw in Ray's dashboard let a malicious website run code on a developer's machine through the unauthenticated jobs API, and CISA lists it as exploited.
Read the analysis →Nullify's AI harness finds, proves and fixes these classes of vulnerability in your code and dependencies, and hands your team a merge-ready pull request.