What problem does it solve? Managing Rainbond platform plugins and feature extensions requires careful state verification, approval gating, and long-running task monitoring, where mistakes like replaying write requests or misreading install states can cause failures or duplicate operations. ## Core Features & Use Cases - Plugin Discovery & State Verification: Lists platform plugins via the Rainbond Console and only treats installed=true with status=RUNNING as ready, distinguishing unavailable states from absent ones. - Safe Lifecycle Operations: Installs, upgrades, and uninstalls plugins through approval-gated write calls, reusing rainbond_upgrade_app and rainbond_delete_app instead of inventing plugin-specific tools. - Long-Task Monitoring Modes: Lets users choose continuous polling until a terminal state or deferred acceptance with read-only status checks on the next wake. - Use Case: When an AI assistant needs the AI Engine plugin before deploying models, this Skill installs the plugin, waits for RUNNING status, then resumes the original model task. ## Quick Start Ask the agent to install the AI Engine plugin on your Rainbond platform and wait until it reaches RUNNING status.