Claude is moving from legal drafts into the files law firms must protect
A law-firm document room where sealed case files connect through controlled permission cables to a calm AI review console, with visible audit checkmarks and no readable legal text.📷 AI-generated image / TECH&SPACE
- ★Anthropic is adding 12 legal plugins and more than 20 MCP connectors for Claude
- ★The integrations target contract law, employment law, litigation and existing legal-tech tools
- ★The main risks are permissions, confidentiality, audit trails and output approval
Legal AI is no longer only about whether a model can draft a contract; it is about how it enters the tools a law firm already uses. The Decoder's report establishes the story, but the useful question is what actually changes behind the announcement.
The new Claude plugins target contract law, employment law and litigation, with links to services such as Thomson Reuters CoCounsel Legal, DocuSign, Everlaw, Box and Harvey. Model Context Protocol documentation helps separate the concrete product, program or research track from plain marketing, while Thomson Reuters CoCounsel Legal supplies the wider context a short news hit cannot carry.
Anthropic is adding legal plugins and MCP connectors for Claude, linking contract law, litigation and tools such as CoCounsel, Harvey and Everlaw.
Close-up of a contract redline workflow showing source-document cards, permission locks and citation markers feeding into an AI workspace.📷 AI-generated image / TECH&SPACE
That changes the risk profile. Hallucinations still matter, but a connector that reaches documents, transcripts and client material raises permission, audit-trail and confidentiality questions. Legal AI becomes infrastructure, not just a text window.
Adoption will depend on how clearly a firm can see what Claude read, which sources it cites and who approves the output. Without that, the plugin becomes a fast way to turn confidential work into a hard-to-explain workflow.
