The fastest way to learn Claude Code is to ask Claude itself: how do hooks work, then have it write one, then have it convert a release checklist into a skill. Claude knows its own capabilities better than any external tutorial.
OSS Scanner gives critical open source projects free, recurring security scans using Anthropic's strongest models. Each finding comes with a proof-of-concept exploit and, where possible, a suggested patch.
In its first week, OSS Scanner reported 77 critical and 803 high-severity findings to open source maintainers. Among them: a high-severity use-after-free memory bug in curl, described by the curl maintainer as one of the project's worst ever vulnerabilities.
Maintainers of critical open source projects can apply through a linked form to receive free periodic vulnerability scans from OSS Scanner.
When building hooks in Claude Code, Claude checks event names against actual documentation rather than relying on its own assumptions, reducing hallucinated hook configurations.
To create a hook in Claude Code, you specify what type you want and Claude builds it for you, rather than requiring you to configure it manually from documentation.
Anthropic is exploring new features for the Claude chat experience and is looking for early testers. A team member's eyes emoji reply suggests something worth watching.
A new Claude feature is rolling out via waitlist, with Pro and Max plan subscribers getting access first and Team and Enterprise plans following. Users are encouraged to sign up through the linked form.
A user flagged several UX issues with the new sections feature: adding a section from an existing group sorts incorrectly, dragging is broken, section names appear smaller and fainter than chat names, chats aren't indented under sections, and threaded chats can't be folded. A team member confirmed the feedback was passed along.
