Urgent.News

What's breaking now, across thousands of outlets.

Tech

CVE-2026-119204: CVE-2026-119204: Uncontrolled Resource Consumption in pypdf Page Tree Flattening

CVE-2026-119204: Uncontrolled Resource Consumption in pypdf Page Tree Flattening Vulnerability ID: CVE-2026-119204 CVSS Score: 7.1 Published: 2026-10-30 CVE-2026-119204 (also tracked as GHSA-cggf-277g-3q54) is a Denial of Service (DoS) vulnerability in pypdf prior to version 6.21.4. The vulnerability stems from uncontrolled resource consumption (CWE-400) during the flattening of a PDF document's…

CVE-2026-119204 is a Denial of Service vulnerability affecting earlier versions of the pypdf library, specifically those before version 6.21.4. The issue arises from an uncontrolled resource consumption (CWE-400) during the flattening of a PDF document's page tree. By supplying a specially crafted PDF file with a hierarchical amplification structure of shared /Pages and /Page nodes, an attacker can cause pypdf to consume excessive memory, leading to a heap memory exhaustion and potential process termination.

The vulnerability is due to the lack of a global budget for copied dictionary entries when flattening the page tree. By referencing shared page nodes multiple times, an attacker can trigger an exponential increase in memory allocation, ultimately resulting in a denial of service. The CVSS score for this vulnerability is 7.1, indicating a high severity.

To mitigate the risk, users are advised to upgrade to pypdf version 6.21.4 or later, which introduces a page_tree_maximum_page_entries limit to prevent excessive memory usage during page tree flattening. This limit is set to 2,000,000 by default, but can be adjusted in memory-constrained environments. Application developers should also catch and handle pypdf.errors.LimitReachedError and pypdf.errors.PdfReadError exceptions in their code when parsing untrusted PDF streams.

To further reduce the risk, operating system and container memory limits, such as RLIMIT_AS and memory cgroups, can be enforced for PDF parsing tasks. Isolating these tasks in dedicated worker processes with strict memory limits and execution timeouts can also help prevent potential exploits.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in Tech

CI/CD for Monorepos 2026: Build Only What Changed

TL;DR: A monorepo with naive CI rebuilds the world on every commit, so pipelines get slower exactly as the repo succeeds.

  • Naive CI rebuilds entire codebase on every commit, causing slowdown
  • Affected-only delivery detects changes, expands through dependency graph
  • Start with path filters in GitHub Actions, graduate to build tools like Nx

Honest talk: motivation is unreliable, so I stopped depending on it

Some days I'm excited to code. Some days I'd rather scroll. If my output depended on mood, I'd ship very little. This post is about building a routine that works on both kinds of day.

  • Motivation fluctuates and routines are more reliable for productivity
  • Define smallest task night before to minimize decision-making
  • Count non-zero work days to emphasize consistency over intensity

SMS OTP Resend in Node.js: Message IDs, Backoff, and Delivery Errors

Short answer: treat an OTP resend as a state transition on the original SMS message, not as a fresh “send again” button. Keep the message ID from the first send, call the resend operation with that…

  • Retain original message ID for SMS OTP resend in Node.js
  • Implement state transitions (pending, accepted, failed) based on response
  • Limit resend attempts to prevent uncontrolled code streams

A Thought Becomes a Phone Call: Neuralink Wires a Grok Agent Into the Brain

One-line summary: A man who has lived with ALS for seven years used a Neuralink implant to summon a Grok agent by mentally saying "Hey Neuralink," then placed a call to his wife using thought alone.

  • Kenneth, ALS patient, controls AI with thoughts via Neuralink
  • First successful thought-to-phone-call conversion demonstrated
  • Breakthrough raises ethical questions on neural privacy and rights

More from Saturday 31 October →