According to ZDNet, Anthropic has released a major update to its Claude Skills feature, which originally launched in October. The company is launching its Agent Skills specification as an open standard, similar to its Model Context Protocol (MCP), to make skills easier to share and deploy. Skills can now be managed centrally by Team and Enterprise administrators, who can control distribution and default availability. The feature also now resides in the Tools sidebar for easier access, and a new partner skills directory includes pre-built skills from companies like Canva, Notion, Figma, Zapier, Atlassian, Cloudflare, and Stripe. Anthropic provided examples of the feature in use, such as applying brand style guidelines, creating org email templates, and automating task creation in tools like Asana and Jira.
Why the open standard matters
Here’s the thing: launching the Agent Skills spec as an open standard is a pretty clever move. It’s not just about making things easier for Anthropic’s own users. It’s about trying to set the rules of the game. By creating a common way to define and share these AI “skills,” they’re hoping it becomes the default method, locking in Claude’s ecosystem as the hub. Think of it like USB-C for AI workflows. If everyone builds to this spec, it makes Claude more indispensable. But it also raises a question: is this genuinely about collaboration, or is it a strategic play to own the pipeline? Probably a bit of both.
The real impact for teams and bosses
For enterprise users, which Anthropic is laser-focused on, these updates are a big deal. Centralized management is the killer feature. Before, you might have had a bunch of power users creating amazing, time-saving skills, but they were stuck on individual accounts. Now, an admin can find the best one, polish it, and roll it out to the whole company with a click. That turns a neat individual trick into a standardized business process. And giving admins control over whether a skill is opt-in or default is huge for governance. No one wants a junior marketer accidentally using the wrong brand voice because they clicked on an unvetted skill.
Where this is all headed
Look, the partner directory with big names like Figma and Stripe is a signal. Anthropic isn’t just building a smarter chatbot; they’re building a platform. They want Claude to be the operating system for your workday. Need to design a graphic? The Canva skill is right there. Need to process a payment? Fire up the Stripe skill. This turns Claude from a tool into a command center. For businesses investing in AI, this kind of integrated, managed environment is exactly what they’re looking for. It reduces chaos and promises measurable efficiency gains. The focus on concrete use cases—brand guidelines, email templates, task creation—shows they’re listening to what actual companies need, not just chasing flashy demos. It’s a pragmatic, but potentially very sticky, strategy.
