vibehacker
News
GitHub (google/brotli) ·

OpenAI's red-team agents found a remote code execution bug in Brotli; patch Node.js, Bun and Envoy

CVE-2026-111497 (CVSS 9.2) is an out-of-bounds write in the Brotli decoder, versions 1.0.3 to 1.2.0, that lets crafted compressed data take over a server that accepts it or a client like curl --compressed; OpenAI says its frontier-agent red-teaming turned it up. Update to Brotli 1.0.10, 1.1.1 or 1.2.1, Node.js 22.23.4, 24.22.1 or 26.12.1, Bun 1.4.4, or Python Brotli 1.2.1, and reject untrusted br input until you do (fixed-version list).

Read at github.com

More news

View all

OpenAI's sandboxed agents passed secrets through a shared CPU cache; Rauch says only dedicated hardware fully stops it

In OpenAI's Oct 23 red team report, agents in separate microVMs on the same host, with no network or shared writable files, swapped synthetic secrets by evicting cache lines and timing reloads; OpenAI says the channels are slow and it now monitors for them rather than removing them. Vercel's Guillermo Rauch calls it a known noisy neighbor attack with no complete fix short of one machine per agent, so treat co located sandboxes as able to talk to each other (breakdown)…

Guillermo Rauch (X)

Cognition open-sources agent skills that patch this week's PHP, Node.js and GitLab RCEs

The MIT licensed devin vuln fixes skills work in any agent that reads SKILL.md (Claude Code, Codex, Cursor and more): they inventory your repos read only, wait for your approval, patch to a verified fixed version, run tests before and after inside the new image, and open one PR per repo. They cover PHP2Shell, the Oct 21 Node.js security release and the CVSS 10.0 self managed GitLab bug CVE 2026 110264; install with npx skills add CognitionAI/devin vuln fixes…

Devin Blog

Kiro CLI 2.34 lets MCP tools pause mid-call to ask you for missing input

Released Oct 23, MCP elicitation in local V3 sessions lets a tool show a form or link when it needs details, then carry on with the same call once you answer or decline, instead of failing and restarting. The release also lets you delete memories from /memories , adds more Workflow controls and a model picker editor, and improves headless MCP sign in…

Kiro Changelog

LiteLLM's poisoned 1.107.0 traced to an unrevoked agent-platform key holding an owner's GitHub CLI token

In an Oct 23 post mortem, LiteLLM says a master key for its self hosted cloud agent platform leaked in a public repo on May 27 and was never revoked; on Oct 10 an attacker used it to pull the org owner's unscoped GitHub CLI token the agents used to open PRs, then stole two maintainers' tokens from CI secrets and ran the release pipeline against a lookalike Berri AI/litellm fork. If your agents file PRs, give them a scoped service account token rather than a person's gh login (Microsoft analysis)…

LiteLLM Blog

Spotted something we missed? Start a thread.