Anthropic open‑sources 80+ specialized legal AI agents on GitHub

- Anthropic published its Claude for Legal repository on GitHub on May 12, offering open-source legal workflow plugins, agent templates and deployment documentation. (github.com) - The public repo lists legal tools for contract review, NDA triage and compliance work, and GitHub showed about 3,400 stars and 539 forks. (github.com) - Anthropic scheduled a May 15 webinar, “How legal teams put Claude to work,” as the next public event tied to the rollout. (anthropic.com)

Anthropic published a public GitHub repository for “Claude for Legal” on May 12, adding an open-source package of legal workflow plugins, reference agents and deployment guides to its broader push into industry-specific AI products. The repository is listed under Anthropic’s GitHub organization and is described as covering legal work in commercial, privacy, product, corporate, employment, litigation, regulatory, AI-governance and intellectual-property teams, as well as law school clinics and students. (github.com) GitHub showed the repository with about 3,400 stars and 539 forks when viewed on May 15. (anthropic.com) Anthropic’s documentation says the package includes “reference agents, skills, and data connectors” and points new users to a quick-start guide for installation. The release adds to Anthropic’s recent sector-specific product launches. Anthropic’s newsroom listed “Agents for financial services” on May 4 and “Introducing Claude for Small Business” on May 13, alongside other product announcements this month. ### Which legal tasks does the software actually cover? (github.com) Anthropic’s legal plugin page says the tools are designed to speed up contract review, NDA triage, compliance workflows, legal briefings and templated responses for in-house legal teams. The page says the software is built for commercial counsel, product counsel, privacy and compliance teams, and litigation support. (github.com) The plugin documentation lists commands for clause-by-clause contract review, NDA prescreening, vendor agreement checks, brief generation and templated responses for requests such as data-subject inquiries and discovery holds. Anthropic says users can configure those workflows to match an organization’s internal playbook and risk tolerances. (anthropic.com) ### Where does the “80-plus agents” claim come from? Anthropic’s managed-agent cookbook says every agent in the repository is shipped in two forms: as a Claude Code plugin and as a Claude Managed Agent template that can be deployed behind a company’s own workflow engine. (claude.com) The cookbook describes those templates as starting points rather than finished products and says they require adaptation to a customer’s document systems, repositories, Slack workspace and review processes. The public GitHub listing itself does not spell out the total in the snippet visible through search results, but third-party coverage published this week described the release as including more than 80 specialized agents, 12 practice-area plugins and more than 20 connectors. (claude.com) Reuters could verify the repository, the practice-area directories and the managed-agent documentation directly on GitHub, but the exact “80-plus” count was not independently enumerated from the visible snippets alone. ### How are those agents meant to be deployed? (github.com) The managed-agent cookbook says templates can be uploaded with a deployment script that creates workers and posts configurations to Anthropic’s agents API. Anthropic’s documentation says each template includes steering examples and an agent-specific README covering security tier and handoffs. The cookbook names example agents including “reg-monitor,” “renewal-watcher,” “diligence-grid,” “launch-radar” and “docket-watcher.” Anthropic says those templates watch sources such as regulatory feeds, contract repositories, virtual data rooms, product trackers and court dockets, with different worker roles handling reading, filtering, deadline mapping and writing. (pasqualepillitteri.it) ### What limits did Anthropic put around legal use? Anthropic’s plugin page says outputs “should be reviewed by licensed attorneys.” A related legal plugin repository in Anthropic’s broader knowledge-work tools says the software assists with legal workflows but does not provide legal advice and that AI-generated analysis should be reviewed by qualified legal professionals before being relied on for legal decisions. (github.com) Those limits appear alongside Anthropic’s push to make the tools easier to install. The GitHub repository is published under an Apache-2.0 license, and the quick-start language on the repo says users can install it in “60 seconds.” (github.com) ### What comes next from Anthropic on legal? Anthropic scheduled a webinar titled “How legal teams put Claude to work” for May 15 at 10:00 a.m. Pacific time. The event page says the session will cover contract review, redlining, extraction and drafting, and will discuss how in-house legal teams and firms have been using Claude Cowork in recent months. (anthropic.com) (github.com) (claude.com)

Get your own daily briefing

Scout delivers personalized news, insights, and conversations tailored to your role and industry.

Download on the App Store

Shared from Scout - Be the smartest in the room.