Anthropic is releasing its latest technology for businesses as an open standard. This means that the company is giving away the blueprints and code for a feature called Agent Skills. These skills are sets of instructions that help the Claude AI perform specific jobs consistently, such as analyzing financial data or writing computer code. By making these tools open source, Anthropic is inviting other companies and developers to use and improve them. This is a strategic move to ensure that their way of building AI becomes the standard used across the entire industry.
This development follows the release of another tool called the Model Context Protocol. While that protocol helps AI connect to external data safely, the new Agent Skills provide the actual knowledge of how to use those tools. Together, they allow an AI to act more like a human employee that can handle complex workflows without needing a person to give it a new prompt for every single step. Major companies like Microsoft are already using these tools in their own software, which shows that Anthropic’s open approach is gaining a lot of support very quickly.
Competing with kindness is the way to go for Anthropic
Anthropic is taking a different path than some of its rivals by choosing to be more open with its technology. While other companies often keep their best business tools behind a paywall or secret code, Anthropic believes that collaboration will lead to faster growth. If every developer uses the same standard for AI agents, it becomes much easier for different AI programs to work together. This creates a massive network where Anthropic’s technology is at the very center. It is a way to compete with giants like OpenAI and Google by making themselves indispensable to the people building the software.
For businesses, this open source approach is very helpful because it prevents them from being locked into just one provider. If a company builds their AI systems using Anthropic’s open standards, they have more freedom to move their data or change their tools in the future. It also means that the technology is being tested and fixed by thousands of people around the world, which usually leads to fewer errors and better security. As 2026 continues, this focus on open standards could help Anthropic win over more professional users who value transparency and flexibility in their digital tools.
Release and pricing information
Agent Skills are now available across the Claude ecosystem, including the web app, mobile applications, and the Claude Code developer tool. Individual users on the Pro and Max plans can access prebuilt skills and create their own custom versions. For developers, the feature is accessible through the Claude API via a dedicated skills endpoint. Anthropic has also launched a directory featuring specialized skills from partners such as Atlassian, Canva, and Notion to help teams automate common software tasks.
The pricing for these features depends on the specific subscription plan or API usage. The Claude Pro plan is priced at $20 per month, while the more powerful Claude Max tiers range from $100 to $200 per month. For businesses, the Team plan costs $30 per user each month with a minimum requirement of five seats. Enterprise customers must contact Anthropic directly for custom pricing, which typically includes advanced security and much larger data limits. Developers using the API for these tools pay based on the number of tokens processed, with the flagship Sonnet 4.5 model costing $3 per million input tokens and $15 per million output tokens.
