Ruby has spent twenty years selling the same promise Java sells: the language manages its own memory, so you never have to count bytes. On November 18, 2026, that promise cracked in the dumbest possible place, in the jump of one letter to uppercase. CVE-2026-132825, now called Ruby2Shell, is a one-byte heap overflow in Ruby’s case conversion that lets an attacker run code on your servers with no password and no account.
Here is the bug, and it is small enough to hold in your head. The German letter ß has no uppercase form in the old single-byte encodings, so Ruby turns it into SS, two characters. The loop that does the conversion checks there is room for one byte and then writes two. Every ß adds one extra byte to the result, and a string with 21 of them overflows the buffer by exactly one byte. That byte writes an S just past the reserved space and corrupts whatever memory sits next to it. This fires on single-byte encodings like ISO-8859-1 and Windows-1252. UTF-8 strings take a different path and are not affected.
What makes this a real story is not the byte. It is that you can reach it without doing anything unusual. The methods upcase, swapcase, downcase with the :fold option, and casecmp? trigger the vulnerable path. In a Rails or Sinatra app running its default config, the framework processes a crafted request and changes the case of one field before your own code ever runs. The attacker does not need to be authenticated, and your application does not need to do anything special. One carefully built HTTP request is enough.
Red Hat rates it 9.8 out of 10 on CVSS 3.1, critical. As of now there is no confirmed exploitation in production, but proofs of concept exist that demonstrate code execution. The researchers at Veria Labs, who gave the flaw its Ruby2Shell name, said they will publish their proof of concept on November 25, 2026, to give everyone time to patch. After that, expect attempts to start.
The fix shipped on November 18: Ruby 4.0.10, 3.4.13 and 3.3.14, announced by the maintainers themselves in their security advisory. The affected ranges are 4.0.0 through 4.0.9, 3.4.0 through 3.4.12, and 3.3.13 and earlier. Ruby 3.2 and older lines have reached end of life and will get no fix. One detail you do not want to miss: updating your gems or your framework does not replace the interpreter. What matters is the Ruby running on your application servers, your background workers and inside your container images, and you have to update and restart each one of them.
The bug also rides along in self-hosted software that bundles its own Ruby. GitHub Enterprise Server is still vulnerable, and the published analysis shows no fixed release yet. GitLab, Mastodon, OpenProject, Chatwoot and Discourse have all shipped patches. In the GitHub advisory, the Ruby maintainers credit Ada Logics and Anthropic for the find. Veria Labs found it independently and was the first to demonstrate remote code execution. Two different teams, one conclusion: the flaw is real, and the hurry to fix it is justified.
If you cannot update today, there is no complete mitigation. What Red Hat and AWS recommend is shrinking your exposure: put internet-facing Ruby apps behind a VPN or an allowlist, and set your WAF or proxy to reject requests that declare ISO-8859 or Windows-125 encodings. Recycling your workers on a schedule limits the damage of a corruption, but it buys time. It does not replace the update.
The pattern is the same one we keep repeating, and that is what stings: a memory bug in the core of the language turns any Ruby app into attack surface. The difference between a hacker who finds flaws and a white hat who reports them is what they do with the discovery. Your to-do list is concrete: inventory the Ruby versions you run, update them to 4.0.10, 3.4.13 or 3.3.14, and restart the services. Leave it for later and you are up against a proof of concept that publishes next week. Until then, least privilege shrinks what an overflow like this can reach.