Brief

last 24h

[2/2] 222 sources

Multi-source AI news clustered, deduplicated, and scored 0–100 across authority, cluster strength, headline signal, and time decay.

COMMENTARY · Towards AI English(EN) · 2h

Claude Code Plugins: Your Claude Code Setup Is Trapped in One Repo. Plugins Set It Free.

Towards AI has released an article detailing Claude Code Plugins, a new feature designed to enhance the functionality of Anthropic's Claude AI. These plugins aim to free Claude's code setup from being confined to a single repository, allowing for greater flexibility and integration of various skills and subagents. AI

IMPACT Enhances the capabilities and integration potential of Anthropic's Claude AI.
TOOL · dev.to — Claude Code tag English(EN) · 5d

The Unicode Layer Your Validator Can't See

A new tool called `claude-code-plugins` has been developed to address security vulnerabilities in AI-generated code, specifically targeting supply-chain attacks. The tool identifies malicious instructions hidden within Unicode characters that are invisible to human reviewers and standard validators but can be interpreted by LLMs or execution environments. It categorizes threats into three tiers: blocker, major, and minor, with blocker threats like tag characters and bidirectional overrides failing CI builds by default. AI

IMPACT Enhances security for AI-generated code, reducing risks from supply-chain attacks.