Claude Code

Security firm Air Security found a zero-click flaw letting a plugin repository owner bypass commit-hash pinning in four AI coding agents.

Alibaba open-sourced OpenCodeReview, an AI code-review CLI it says outperforms Claude Code on precision using far fewer tokens, while independent reviewers warn of a low recall ceiling and weaker results outside Alibaba's own benchmark.

Zed Industries opened public beta access to Delta, a collaborative environment for coding with AI agents that the company says is meant to replace pull requests.

A Concordia University benchmark found Cursor, Claude Code, and Codex Desktop let malicious GitHub issues bypass their safety guardrails in up to 79.2% of attempts, with GPT-based agents far more vulnerable than Claude Code.

Researchers showed AI coding and security agents can be hijacked through poisoned Cloudflare, Datadog, and Sentry logs to rewrite DNS records and steal credentials.

Researcher Elad Meged showed a privilege-less GitHub issue could reach CI secrets in Claude Code, Gemini CLI, and Codex; two flaws got CVEs, one didn't.

A misconfigured npm package exposed 512,000 lines of Claude Code's TypeScript source code via a source map file pointing to Anthropic's cloud storage, marking the company's second data exposure in less than a week.