AICoder

AI Coding Live News

First-hand AI coding updates from 𝕏/Twitter frontier labs and researchers, distilled by AI with entity benchmarks.

Explore live on 𝕏 ↗

AI Coding Live News

A
Anthropic Frontier Red Team@AnthropicAI·
👀65
🔥86

Anthropic's Frontier Red Team details five exploit chains an unreleased internal Claude model built for Chrome, Safari, Android, Windows and Linux; Chrome 156 fixes 62 CVEs

Anthropic's Frontier Red Team published a research post on Oct 27 saying it spent the past few months using an unreleased internal research model to find critical vulnerabilities in browsers and operating systems, and that after initial prompting Claude wrote five full exploit chains: one-click remote code execution with a sandbox escape in Chrome, a same-origin policy bypass in Safari, data extraction from a locked Android phone over USB, a remote denial of service against Windows HTTP/3 (QUIC), and zero-click root code execution on Linux servers with a multipath TCP listener. Anthropic says Google assigned 62 CVEs to its reports, all critical or high, fixed as of Chrome 156, and every issue went through coordinated disclosure and has been fixed. The model's name, the amount of human involvement and the compute cost were not disclosed.

Anthropic's Frontier Red Team details five exploit chains an unreleased internal Claude model built for Chrome, Safari, Android, Windows and Linux; Chrome 156 fixes 62 CVEs

Key Takeaways

  • The Chrome chain links 4 bugs: V8 heap corruption, a V8 sandbox escape that defeats ASLR, a CacheStorage race condition and a FedCM use-after-free; the PoC launches Calculator within seconds of a page loading in default Chrome on macOS
  • Per Anthropic, Google assigned 62 CVEs to its reports, all critical or high, including 13 sandbox escapes and 21 renderer-compromise bugs, fixed as of Chrome 156
  • The Windows kernel QUIC out-of-bounds read is CVE-2026-33096, with likely several million exposed servers by Anthropic's estimate; the Linux MPTCP zero-click root bug is CVE-2026-98432, where the PoC leaks a kernel pointer in about 48 minutes and reaches root command execution about 5 minutes later
  • Method: many parallel scanning agents scored by the extra vulnerabilities they add, with coverage tracking and automated variant analysis on every confirmed finding; Anthropic recommends scanning at PR time and has released a Claude Code security plugin
  • The model is an unnamed internal research model; Anthropic did not publish the share of human work, failed attempts or compute cost
Read full story
Linked Models & Agents:
⚡ Claude (unreleased internal research model)🤖 Claude Code🌐 Anthropic🌐 Google🌐 Apple🌐 Microsoft
#Anthropic#Claude#Security#vulnerability research#Chrome#Safari#Linux#Frontier Red Team
O
OpenAIOpenAI·
👀51
🔥84
View on 𝕏 ↗

OpenAI red-team agents find a remote code execution bug in Brotli (CVE-2026-111497); Node.js 22, 24 and 26 are patched and developers should upgrade

On 2026-10-23 (22:55 UTC), OpenAI disclosed that during red-teaming of its frontier model training environment, agents powered by its most cyber-capable models found an unauthenticated remote code execution bug in the Brotli compression library. Brotli ships inside Node.js, Bun, Envoy and much more: an unpatched server that accepts compressed input, or a client such as curl --compressed that accepts compressed responses, could be taken over. Brotli released v1.2.1, v1.1.1 and v1.0.10 on Oct 21 with the fix (CVE-2026-111497), and Node.js shipped 22.23.4, 24.22.1 and 26.12.1 the same day, crediting OpenAI Codex Security.

Key Takeaways

  • CVE-2026-111497: an out-of-bounds write in the Brotli decoder that permits remote code execution (Brotli v1.2.1 release notes)
  • Brotli shipped three fixed releases on 2026-10-21: v1.2.1, v1.1.1 and v1.0.10
  • Node.js 22.x, 24.x and 26.x are affected; fixed in 22.23.4, 24.22.1 and 26.12.1, rated high by Node.js
  • The root cause was fixed in Brotli source in February, but nobody knew it was a security bug and no release followed, so downstream never picked it up
  • Found by agents OpenAI used to red-team its training environment; the Node.js advisory credits OpenAI Codex Security
Read full story
Linked Models & Agents:
🤖 OpenAI Codex Security🌐 OpenAI🌐 Google🌐 Node.js
#Brotli#CVE-2026-111497#Node.js#Security#OpenAI#Codex Security#red teaming
O
OpenAIOpenAI·
👀66
🔥90
View on 𝕏 ↗

OpenAI Patch the Planet: GPT-5.6 Sol helped find two critical PHP RCE flaws, fixed in PHP 8.5.12, 8.4.27, 8.3.36 and 8.2.35

On Oct 21, PHP published a security advisory fixing two critical unauthenticated remote code execution flaws (CVE-2026-109618, CVE-2026-109619) affecting every PHP install serving web requests, with mod_php at highest risk. OpenAI disclosed the same day that Calif found a flaw using GPT-5.6 Sol under its Patch the Planet initiative, with Wiz independently finding and reporting both; OpenAI coordinated a two-month disclosure and committed $1B in subsidized Daybreak access for under-resourced defenders.

OpenAI Patch the Planet: GPT-5.6 Sol helped find two critical PHP RCE flaws, fixed in PHP 8.5.12, 8.4.27, 8.3.36 and 8.2.35

Key Takeaways

  • Flaws: CVE-2026-109618 (memory corruption via OOM handling) and CVE-2026-109619 (stale global state across requests), both unauthenticated RCE (PHP official)
  • Fix: upgrade to PHP 8.5.12, 8.4.27, 8.3.36 or 8.2.35; PHP says there are no complete mitigations and updating is the only safe choice
  • Scope: PHP-FPM, Apache mod_php, php-cgi, FrankenPHP and LiteSpeed are affected, with a demonstrated exploit on mod_php
  • AI role: Calif found a flaw with GPT-5.6 Sol, Wiz reported both independently; OpenAI coordinated a ~2-month disclosure (vendor account)
  • Investment: OpenAI committed $1B in subsidized Daybreak access and will ship a Codex skill to check PHP versions
Read full story
Linked Models & Agents:
⚡ GPT-5.6 Sol🤖 codex🤖 daybreak🌐 OpenAI🌐 PHP🌐 Calif🌐 Wiz🌐 Cloudflare
#Security#PHP#CVE#RCE#Patch-the-Planet#AI-vulnerability-discovery#Codex
N
NUS School of Computing@NUSingapore·
👀115
🔥97

Last Secure Code Benchmark: Vibe Coding Under Fire as Top Agents Produce Secure Code on Only 25.1% of Tasks

While AI coding agents and 'vibe coding' workflows synthesize an unprecedented volume of production software, code security remains highly fragile. Researchers from NUS and Monash University introduced the Last Secure Code Benchmark, evaluating 450 tasks derived from 150 real CVE vulnerabilities across 7 languages (Java, C, C++, JS, Python, PHP, Go) and 24 CWE categories (including 21 of the 2025 CWE Top 25 and 8 of the Top 10 KEVs). While frontier agents achieve 73.6% to 91.3% functional completion rates, even the strongest model (Claude Fable 5) produces code that is both functional and secure on only 25.1% of tasks, exposing acute supply-chain risks in autonomous vibe coding.

Last Secure Code Benchmark: Vibe Coding Under Fire as Top Agents Produce Secure Code on Only 25.1% of Tasks

Key Takeaways

  • First comprehensive vibe-coding security benchmark: 450 tasks based on 150 CVEs spanning 7 languages, covering 21 of the 2025 CWE Top 25 and 8 Top 10 KEVs
  • The functionality illusion: while agents pass functional specs at 73.6%–91.3%, dual functional-and-secure pass rates collapse to 16.4%–25.1%
  • Vulnerability defense crisis: across full-repository and file generation scenarios, top models resolve only 12% of vulnerability environments securely
Read full story
Linked Models & Agents:
⚡ Claude Fable 5⚡ GPT-5⚡ Claude 3.5 Sonnet⚡ DeepSeek-Coder-V2🤖 Coding Agents🤖 Autonomous Software Engineering🌐 National University of Singapore🌐 Monash University
#Last Secure Code#Vibe Coding#Coding Agents#Cybersecurity#CWE#Software Vulnerability
ADSponsored
A
Anthropic@AnthropicAI·
👀104
🔥92
View on 𝕏 ↗

Anthropic Cyber Mission: free OSS Scanner reports 77 critical and 803 high findings in its first week, plus Critical Infrastructure Defense with 11 partners

On Oct 8 Anthropic launched the Cyber Mission: (1) free opt-in OSS Scanner for critical open-source projects using strongest models including Claude Mythos, with PoC/explanation/candidate patches and >90% expected true-positive rate, no human triage; (2) Critical Infrastructure Defense Program with 11 founding partners including CrowdStrike, Palo Alto, Dragos, and Rockwell. Project Glasswing has surfaced 29,000+ candidate vulns. Oct 19 update: first scans have reported 77 critical and 803 high model-rated findings to maintainers; curl 8.23.0's high-severity HTTP/2 use-after-free CVE-2026-92392 was found by an early version.

Key Takeaways

  • Two tracks: free opt-in OSS Scanner (model-only reports) + Critical Infrastructure Defense Program (OT/power/water/transport; 11 founding partners)
  • Glasswing legacy: 29,000+ candidate vulns, ~6,000 human-triaged; ~5,000 reports sent when maintainers asked for bulk
  • Quality bar: 88% of 97 critical/high samples met CVD bar; expected true-positive >90% (no human review; severity can be off)
  • Enroll: core maintainers PR into Anthropic’s enrollment repo; also Claude for OSS free Max + Cyber Verification Program
  • First week (vendor-reported, model-rated, unreviewed): 77 critical and 803 high findings sent to maintainers; curl 8.23.0 fixed CVE-2026-92392 (HTTP/2 upload use-after-free, 8.12.0–8.22.0) found by an early version
Read full story
Linked Models & Agents:
⚡ Claude Mythos⚡ Claude🌐 Anthropic🌐 CrowdStrike🌐 Palo Alto Networks🌐 Dragos🌐 Rockwell Automation
#Anthropic#Cyber Mission#OSS Scanner#Glasswing#Security#Open source#Curl#CVE-2026-92392
H
Hugging Face Daily Papers@HuggingFace·
👀93
🔥99

CheckerBench: ECNU and Peking University Introduce the First Executable Benchmark for Coding Agents Synthesizing Static-Analysis Checkers

Prevailing coding agent benchmarks evaluate isolated patch synthesis (e.g. SWE-bench) or vulnerability detection, neglecting whether agents can construct reusable, production-grade static-analysis checkers from scratch. Developing static analyzers requires interpreting defect specifications, traversing multi-file AST semantics, writing engine-specific logic, and refining implementations through iterative compilation and diagnostic feedback. Researchers from East China Normal University, SJTU, and Peking University unveil CheckerBench (arXiv:2610.07557), an executable suite comprising 300 tasks derived from 297 CVEs across 167 repositories, 85 CWEs, and five programming languages. Powered by the CheckerLab harness, evaluation across 21 model-harness configurations reveals a mean Pass@1 rate of only 32.30% (peak 45.33%), demonstrating that end-to-end program analysis tool synthesis represents a profound new frontier for software engineering agents.

Key Takeaways

  • ECNU, SJTU, and Peking University launch CheckerBench, the first executable benchmark for coding agents synthesizing static analysis checkers
  • Features 300 tasks spanning 297 CVEs and 85 CWEs across C/C++, Go, Java, Python, and Rust with the CheckerLab framework
  • 21 model-harness setups achieve a mean Pass@1 of only 32.30% (peak 45.33%), exposing gaps in compiler-in-the-loop tool synthesis
Read full story
Linked Models & Agents:
⚡ CheckerBench⚡ CheckerLab⚡ SWE-bench⚡ Coding-Agents🤖 Static-Analysis-Agent🤖 Software-Security-Agent🤖 Checker-Synthesizer🌐 East China Normal University🌐 Peking University🌐 Hugging Face
#CheckerBench#Coding Agents#Static Analysis#Vulnerability Detection#Compiler In The Loop#AI Benchmarks
R
Ren@Ryrenz·
👀70
🔥90
View on 𝕏 ↗

Tencent Zhuque open-sources AI-Infra-Guard: CVE scanning for AI stacks plus MCP/Skill checks

@Ryrenz highlights Tencent Zhuque Lab’s open-source AI-Infra-Guard (github.com/Tencent/AI-Infra-Guard): a full-stack AI red-team/scanner with 146+ AI components and 2000+ CVE rules (a recent drop added 155 rules across 40+ components). Beyond infra such as n8n, vLLM, MLflow, and LangFlow, it scans MCP servers and Agent Skills for prompt-injection and unauthorized file-read risks. Apache-2.0; intended for internal use and currently has no auth gate.

Key Takeaways

  • Scans AI app stacks for known CVEs across components like n8n, vLLM, MLflow, and LangFlow.
  • Also scans MCP servers and Agent Skills for prompt injection and unauthorized file reads.
  • Apache-2.0 from Tencent Zhuque Lab; internal-use positioning with no auth—do not expose publicly.
Read full story
Linked Models & Agents:
🌐 Tencent (腾讯混元)
#AI-Infra-Guard#Tencent#Zhuque#MCP#Agent Skill#CVE#Security#open source
D
Dan Kornas@DanKornas·
👀52
🔥89
View on 𝕏 ↗

Mira open-sources a self-hosted AI code reviewer with full-repo indexing

@DanKornas highlighted miracodeai/mira, a self-hosted AI code reviewer for teams that keep diffs, indexes, review history, and CVE data in their own SQLite or Postgres. It supports OpenRouter and OpenAI-compatible endpoints including vLLM/Ollama, applies per-repo rules via .mira.yaml, and auto-reviews PRs on GitHub, GitLab, and Forgejo/Codeberg. Apache-2.0.

Key Takeaways

  • Self-hosted by default: code and review data stay on your infrastructure.
  • Full-repo indexing plus custom rules go beyond a single diff.
  • Supports GitHub, GitLab, and Forgejo; models via OpenRouter or local OpenAI-compatible endpoints.
Read full story
Linked Models & Agents:
🤖 Claude Code (Anthropic)🤖 Cursor Composer (Agent Mode)
#Mira#AI code review#Self hosted#OpenRouter#Ollama#GitHub#GitLab#Apache-2.0#open source
You've caught up with all stories