Anthropic published a free, open-source set of legal skills for Claude Cowork, prompting stock declines among several US legal-tech companies.

What the skills do

The legal skill pack can analyze contracts for risks and inconsistencies, assemble templated filings, and track case deadlines across workflows.

Integrations include direct connectors to Slack, DocuSign and Lexis+, enabling automated notifications, signature workflows and research links within Claude Cowork.

• Contract analysis for risk identification and inconsistency detection across clauses and obligations.
• Template-based preparation of court and transactional documents with adjustable parameters for jurisdictional differences.
• Deadline and milestone tracking linked to communication channels and signature platforms for operational follow-up.
• Anthropic also published a companion document listing 50+ distinct skills to adapt agents for several legal tasks and jurisdictions.

Market reaction

Public reaction included declines in shares of established legal-tech vendors that sell turnkey integrations and managed workflows to corporate clients.

Those companies position products as preconfigured platforms requiring paid subscriptions, while Anthropic provided comparable ready-to-use workflows for free and with open code.

Analysts compared the move to historical cases where open models disrupted incumbent enterprise suppliers by reducing adoption friction and cost for buyers.

Implications

The release marks a structural shift toward narrow, task-specific AI tools that lower barriers to automation for routine legal workflows.

Within two years, many standard professional SaaS offerings may become obsolete as open stacks and base models host common workplace functions.

Companies that continue to sell proprietary data, specialized expertise or regulatory compliance wrappers will retain commercial value that is harder to automate.

Risks and caveats

Open legal skills reduce manual effort but do not remove the need for human review in legal workflows.

Legal drafting often depends on unique facts and local rules, which remains difficult for models to handle without expert supervision.

Conclusion

Anthropic’s contribution expands access to practical legal automation while highlighting continued demand for professional judgment, specialized data and regulatory assurance.