Ruby Flaw Hits Default Web Apps, Plus 3 Stories
A Ruby security flaw is the lead today because this one isn’t confined to somebody’s exotic setup. In default Ruby on Rails and classic Sinatra applications, a specially crafted web request can reach the vulnerable code before the application checks who sent it. Also in today’s brief: faster access to medical records, a Bell Canada breach with a slippery “agentic” label, and a Google Vault change that closes part of an external-chat retention gap.
The Ruby security flaw reaches default web apps
The Ruby project has released versions 3.3.14, 3.4.13, and 4.0.10 to fix seven memory-safety problems. The most immediately worrying one, now tracked as CVE-2026-132825, is a one-byte heap buffer overflow in Ruby’s string case-conversion code. Ruby says working proof-of-concept attacks can turn some of these flaws into remote code execution. In plain English, an attacker may be able to make the server run the attacker’s commands.
The path into ordinary web applications is what moves this above the usual pile of scary-looking vulnerability numbers. Rails and classic Sinatra use Rack middleware to support a form field called _method. It lets a web form pretend to send PUT, PATCH, or DELETE even though browsers normally submit forms with GET or POST. Rack converts that value to uppercase while routing the request. A crafted value using a legacy character encoding can make Ruby write two bytes where it allowed room for one.
That conversion happens before the request reaches the application’s login check. So “the admin page requires a password” isn’t an answer here. Neither is “we don’t write Ruby,” if the business runs a self-hosted product or container image that bundles it. Cloudflare identified affected releases of products including GitLab, GitHub Enterprise Server, Mastodon, OpenProject, Discourse, and Chatwoot at the time of disclosure. Vendors will have their own updates and rebuild instructions, so check the product advisory rather than guessing from a Ruby version string.
If you operate a Rails or Sinatra site, inventory every runtime, upgrade, rebuild container images, and restart Ruby processes and background workers. Ruby 3.2 and older won’t receive this fix. Cloudflare has enabled blocking rules, including on its free plan, but calls the web application firewall a compensating control. That’s the right phrase. A WAF can’t inspect traffic that bypasses it, internal service calls, queued jobs, email imports, or files that reach the same vulnerable code. It buys patching time. It doesn’t make the bug disappear.
This is the third major language or server-layer issue we’ve covered in a few days, following Java and NGINX updates. The lesson isn’t that every business needs to become a programming-language archaeologist. It is that an asset list stopping at “website” is too shallow. You need to know what runs the website, what is hidden inside its containers, and who is responsible for updating it.
WordPress security & maintenance
Keeping WordPress current is only part of keeping it healthy. Raymond Tec handles updates, backups, security monitoring, compatibility problems, access cleanup, and maintenance — plus the assorted weirdness that accumulates on a site over time.
Medical-record requests get a shorter federal clock
The U.S. Department of Health and Human Services has finalized a substantial update to the HIPAA Privacy Rule. For patients, the clearest change is time: covered entities will generally have no more than 15 calendar days to respond to a request for protected health information, down from 30. They may take one extension of up to 30 days with written notice, but HHS says the records must still be provided as soon as practicable. Fifteen days is a ceiling, not a new excuse to wait until day fifteen.
The rule also clarifies access through personal health apps when a provider can produce electronic records in that form, changes when copies must be free, requires advance notice of fees, and makes it easier to direct an electronic copy to another person. HHS also reduced some identity-verification friction and expanded permitted information sharing for care coordination.
That combination matters because access rights live or die in the workflow. A right on paper is not much help when a patient is bounced between a portal, a records vendor, a fax number, and an unexpected bill. On the provider side, though, a shorter deadline won’t be solved by telling the same understaffed records team to type faster. Practices, health plans, and their vendors need request intake, identity checks, fee notices, exports, and escalation paths that agree with one another.
No one needs to race to do anything this weekend. The final rule becomes effective 60 days after publication, and regulated organizations generally get 180 days after that to comply. Patients should know the stronger right is coming. Providers should use the runway to test the full request process, including the awkward cases, not just update a policy page.
The rules around technology matter too
Platforms, privacy, speech, competition, surveillance, copyright, and regulation increasingly determine what technology companies can build and what the rest of us have to live with. Browse more Raymond Tec News for practical coverage of technology policy and digital rights.
Bell Canada’s “agentic” breach needs less branding, more detail
Bell Canada says an unauthorized party accessed several systems on November 17 and stole data belonging to enterprise customers. The affected information came partly from a preproduction Maximo environment used to order and provision telecom circuits. Bell told customers the data may include business names and contact details, ticket titles, some IP addresses, and limited context about network elements. The company says credentials, highly sensitive personal information, attachments, and PDFs weren’t in the exposed dataset, and customer services weren’t disrupted.
Bell called the intruder an “agentic threat actor,” which strongly suggests automation or AI-assisted tools. But that isn’t an incident report. It doesn’t tell us which agent was used, what it actually did, how the attacker got in, or whether automation materially changed the attack. Until Bell publishes more, “agentic” is an attribute the company assigned, not an explanation readers can evaluate.
The useful part is the stolen context. Business contact details and circuit information aren’t necessarily catastrophic on their own, but they can make a later phone call or email sound authentic. An attacker who knows a provider, ticket subject, IP address, or planned network change has better material for impersonating Bell or an internal technician. Bell says it shared indicators of compromise and threat-hunting information with affected customers and Canada’s cyber agency. Notified businesses should get those indicators into their monitoring, warn the people who handle telecom changes, and verify unusual requests through a known second channel.
Technology is rarely just about the technology
Some of the most important technology stories aren’t product launches at all. They’re about health, privacy, education, law, accessibility, work, and what happens when technology reaches ordinary people. Browse more Raymond Tec News for the stories worth understanding without the hype.
Google Vault will retain more external Chat conversations
Starting in mid-February 2027, Google Vault administrators will be able to retain, hold, search, and export their organization’s copy of messages from external Google Chat spaces and group direct messages. Today, only the organization that created one of those conversations can preserve it in Vault. After the change, each participating organization will keep its own copy from the point its first user joins until its last user leaves.
This closes a real compliance gap, but it has edges. Messages sent, edited, or deleted before anyone from your company joined may not be captured in their earlier form. Conversations with history turned off still won’t be available in Vault. And the policy controlling what users see in the Chat interface may belong to the external organization that created the conversation, while your own policy controls the copy your administrators can search. That split will confuse somebody eventually, so administrators should document it before a legal hold makes the distinction urgent.
Google will enable the behavior automatically for supported editions, with no separate switch. There’s no immediate action today, but businesses using external Chat should review retention rules, legal-hold procedures, and who may join external conversations before February. “It disappeared from the chat window” and “we no longer retain it” are not the same statement.
Still in a reading mood? The Raymond Tec News archive covers security, AI, small-business technology, policy, and the places technology collides with ordinary life — without requiring a computer-science degree to get through it.
Sources / Further Reading
- Ruby: security advisories for Regexp and string case conversion
- Cloudflare: how ruby2shell reaches Rails and Sinatra
- HHS: 2026 HIPAA Privacy Rule final-rule fact sheet
- Federal Register: full HIPAA final rule
- BleepingComputer: Bell Canada breach report and company statement
- Google Workspace: expanded Vault support for external Chat
Photo by Godfrey Atima on Pexels.
