Anthropic has published a GitHub repository called claude-for-legal, offering a collection of plugins that automate common legal workflows. The suite covers in‑house commercial, privacy, product, corporate, employment, litigation, regulatory, AI‑governance and academic use cases. Each plugin ships as a Claude Cowork or Claude Code extension, and can also be accessed through the Claude Managed Agents API. It includes QUICKSTART.md for a 60‑second install guide.

Deployments run from a single system prompt, so the same reasoning engine powers both desktop‑based Cowork sessions and server‑side managed agents. Claude guardrails enforce attorney‑level review: every citation is attributed, privilege defaults stay conservative, jurisdiction assumptions surface, and explicit gates block any output from being filed without human sign‑off. The output remains a draft, not legal advice.

Practitioners can invoke named agents such as Vendor Agreement Reviewer, DSAR Responder or Termination Reviewer with a single command, while connectors integrate Slack, Google Drive, Ironclad and other legal platforms. By automating triage, amendment tracking and compliance calendars, the toolkit accelerates document review without replacing lawyer judgment. Teams that adopt it report faster drafting cycles and more consistent playbook enforcement.