What problem does it solve? Building and operating Miaoda (Spark) applications on Feishu/Lark requires juggling app creation, local or cloud development, databases, file storage, environment variables, releases, observability, roles, and automation triggers. This Skill routes every such intent to the correct lark-cli command with the right safety guardrails. ## Core Features & Use Cases - Full app lifecycle: Create html or full_stack apps, initialize local Git repositories, run cloud AI generation sessions, and publish releases with polling until finished. - Platform resource management: Execute SQL against app databases with platform rules (audit columns, RLS, user_profile), manage file storage, environment variables, access scopes, roles, and OpenAPI keys. - Observability and automation: Query logs, traces, metrics, and PV/UV analytics; configure cron, record-change, webhook, and Feishu approval triggers with strict high-risk confirmation protocols. - Use Case: A user asks to build a polling tool and deploy it. The Skill decides between local and cloud development, creates the app, initializes the repo, commits and pushes to sprint/default, creates a release, and returns the online URL. ## Quick Start Ask the assistant to create a new Miaoda app named "team-polls", develop it locally, and deploy it to get a shareable link.