Feishu Interactive Cards

**Judgement** A solid, production-leaning skill for building interactive Feishu (Lark) workflows in OpenClaw. It is especially suitable whenever an agent would otherwise make a risky or ambiguous decision based on free-text input. **What it does well** - Enforces a clear principle: whenever there is uncertainty, use an interactive card instead of plain text. - Provides ready-made templates for confirmations, todos, polls, and forms, plus custom templates. - Includes a long-polling callback server that auto-reconnects and integrates with the OpenClaw Gateway, so no public IP is needed. - Offers concrete code examples for sending cards, handling callbacks, and updating cards based on user actions. - Covers UX best practices (clear titles, danger buttons for destructive operations, carrying state in button values) and operational patterns (timeouts, deduplication, error handling). **Key risks / limitations** - Requires running and maintaining an additional Node.js callback server and correctly configuring Feishu long-polling and the OpenClaw Gateway, which adds operational complexity. - While security guidance is strong, actual safety depends on how callback handlers are implemented by the user (e.g., correct input validation and permission checks). - Specialized to Feishu; not useful for other channels and may require familiarity with Feishu’s card schema and app configuration. **Recommended scenarios** - Agents that interact heavily with Feishu and frequently need user confirmations (e.g., deleting resources, changing settings, triggering deployments). - Workflows that benefit from structured user input in Feishu: forms, checklists, approvals, polls, or multi-step wizards. - Teams wanting to reduce ambiguity and mis-executions by making users choose from explicit buttons instead of relying on free-text commands.