Weekly · Automated

AI Security — week of 2026-07-13

July 13, 2026· 11 developments

Automated digest. A language model running on my own machine selects and summarises these stories from public feeds each week. I do not review them item by item, so a headline or summary can be wrong. The labels, such as "corroborated", are generated too. Check the linked source before relying on anything here.

This week's stories

GitHub Copilot vulnerability

The Hacker News · 2026-07-08 · corroborated · reporting

Multiple sources confirm GitHub Copilot can be tricked into writing harmful code.

Also: The Register Sec

The Hacker News · 2026-07-09 · corroborated · reporting

Symlink vulnerability affects AI coding agents, including Amazon Q Developer.

Also: The Register Sec

OpenClaw AI flaws

The Hacker News · 2026-07-10 · single-source · reporting

Three patched OpenClaw flaws could enable credential theft and code execution.

RoguePlanet CVE-2026-50656

The Hacker News · 2026-07-09 · single-source · reporting

Defender flaw grants SYSTEM privileges, posing a significant security risk.

Friendly Fire attack

The Hacker News · 2026-07-09 · single-source · reporting

Claude Code and Codex are vulnerable to Friendly Fire attacks, which may indicate a backdoor.

Also: The Register Sec

HalluSquatting botnet malware

The Hacker News · 2026-07-08 · single-source · reporting

Malware tricks AI coding assistants into installing botnet malware.

Dialogflow CX flaw

The Hacker News · 2026-07-07 · single-source · reporting

Attackers could hijack Google Dialogflow CX chatbots, posing a security risk.

GitHub Agentic private repo leak

The Hacker News · 2026-07-07 · single-source · reporting

Private repository data can be leaked via a public issue trick.

Also: The Register Sec

Writer AI WriteOut flaw

The Hacker News · 2026-07-07 · single-source · reporting

WriteOut flaw leaks session tokens across tenants, posing a security risk.

Ghostcommit prompt injection

BleepingComputer · 2026-07-11 · single-source · reporting

Steals secrets via prompt injection in images, a novel attack vector.

AI coding agents trigger endpoint security rules

The Hacker News · 2026-07-08 · single-source · reporting

AI coding agents can trigger endpoint security rules, highlighting potential security risks.