[Feature Request] Add programmatic trigger mechanism for Claude Design sessions via API or webhooks

Bug Description
Feature request: programmatic bridge between Claude Code and Claude Design (claude.ai/design) I run a workflow where Claude Code and Claude Design collaborate on the same design project. Claude Code can already read/write files in the design project through the DesignSync MCP tool (list_files, get_file, write_files), which is great: my coding session deposits spec/dispatch files into the project and pulls the generated mockups back without any copy/paste. The missing piece is triggering. Claude Design only runs when I manually open claude.ai/design in a browser and type a message in the design chat. There is no API, webhook, deep link, or routine that can start or continue a design conversation, so the round trip Claude Code → Claude Design → Claude Code always has a human browser step in the middle, even when the prompt is just "read docs/specs/x.md and produce the mockup". What would solve it, in order of preference: A trigger surface for design sessions: an endpoint or routine-style mechanism that starts a design generation on a project with a given instruction (even fire-and-forget, results land as files that MCP can read back). A "watched folder" or on-file-change hook in design projects: new file in a conventioned path → Claude Design processes it. At minimum, a deep link like claude.ai/design/<project>?q=... that pre-fills the prompt, so the manual step becomes one click. Context: all Claude Code automation surfaces (routines, claude-cli:// deep links, Claude Code on the web, Slack) only ever reach Claude Code sessions, and the Compliance API is read/delete only, so today there is no supported path at all. The file-sync half of the bridge already exists and works well; it is only the trigger half that is missing.

Environment Info

• Platform: darwin
• Terminal: vscode
• Version: 2.1.220
• Feedback ID: e0905e52-9cf7-46bd-b80b-357500b13e8f

Errors

[]