Claude for Small Business Passes 900,000 Installs, Anthropic Adds 43 Workflows

Half of owners' biggest worry was data security. Anthropic's answer wasn't "trust us" — it was "the permissions you already set still apply to Claude," turning an abstract trust question into a concrete, checkable rule.

Grace Patel · September 18, 2026

Anthropic announced the latest progress on Claude for Small Business on September 15, 2026: since its May 13 launch, this small-business toolkit built into Claude Cowork has been installed and used more than 900,000 times. The announcement also confirms a significant expansion — workflows grew from the original 15 to 43, connected third-party software expanded from 8 to 27, and the nationwide training tour is restarting.From a Toggle to 43 WorkflowsClaude for Small Business works as a...

news

Cowork Merges Into Claude's Main Interface as Docs and Slides Launch in Beta

You used to have to figure out where a task belonged before you could start....

practice

Fable 5.1's Cache Reads Are 75% Cheaper — How Much You Actually Save Depends on Your Bill's Cache Share

The 25% to 45% Anthropic quotes isn't a range where you get to pick...

practice

Three Breaking Changes When Upgrading to Fable 5.1 — One of Them Can Silently Break Your Agent

The first two breaking changes throw an error. The third one doesn't — it...

Useful Resources

Claude API Status → Model Pricing → Prompt Playground → Token Counter → MCP Servers → LLM Benchmarks → Model Comparison →

Glossary

View All →

Agent Permission Scoping

The practice of deliberately granting an <a href="/en/glossary/core-concepts/ai-agent/">AI Agent</a> only the minimum set of permissions needed for its current task, rather than granting broad access far beyond actual need for convenience, so the potential impact of an error stays as contained as possible.

What is <a href="/en/glossary/agent-permissions/agent-permission-scoping/">Agent Permission Scoping</a>, and how does it differ from the intuition...

advanced

Least-Privilege Agent

Granting an agent only the minimum set of tools and data access strictly needed for its current task, rather than starting broad and narrowing later.

What does the principle of least privilege actually mean when applied to an agent? Least privilege isn't an AI-specific concept — it's a...

intermediate

Prompt Injection Defense

Designing a mechanism so an agent, while processing external data (like webpage content or user-uploaded documents), can distinguish original task instructions from malicious instructions potentially hidden within the data itself, avoiding being misled into performing unintended actions by content buried in that data.

What is prompt injection defense, and how does it differ from typical <a href="/en/glossary/prompt-techniques/jailbreaking/">jailbreaking</a>...

intermediate

Weekly Picks

practice

Claude Code Adds --restricted Mode: A Minimal-Authority Starting Point for Unfamiliar Projects

--restricted isn't an OS sandbox. It's a startup config that defaults to "everything's pulled...

Marcus Ide

practice

Claude Code Fixes `/cd`: New Directory Settings Now Apply Immediately, No More Waiting for Resume

There used to be a gap between moving directories and the new settings actually taking effect —...

Derek Finch

practice

Claude Code Adds a Startup Warning: `Bash(git * main)` Matches Far More Than You'd Expect

`Bash(git * main)` looks like it has two constraints written in, but it's actually looser than a...

Marcus Ide

practice

Claude Code's --worktree Now Accepts a GitLab Merge Request URL Directly, No Number Conversion Needed

One fewer step of "look up the number, then manually assemble the format" sounds small, but if...

Derek Finch