Claude Code adds live artifacts

Anthropic is turning agent work from a private terminal transcript into a collaborative surface, which matters as coding agents move deeper into team workflows.

• –Artifacts use full session context, including the codebase, connectors, skills, and conversation, so teams can inspect outputs without asking the agent operator for a manual walkthrough.
• –The strongest use case is asynchronous engineering review: PR explanations, incident dashboards, architecture maps, and release checklists can stay current at the same private link.
• –Org-only sharing, admin controls, retention policies, and compliance API visibility make this clearly aimed at enterprise teams, not casual artifact sharing.
• –The feature also raises the bar for rival AI coding tools: status reporting and explainability are becoming part of the developer-agent product, not a separate documentation chore.