AI Agent Hub
Back to plugins
⚙️

dsh-pelican

Workflow Updated 2026.08.25

Run the following command in DeepSeek Harness:

dsh plugin install Little-Star888/dsh-pelican

Paste the following prompt into your AI chat to install this plugin:

Run dsh plugin install Little-Star888/dsh-pelican inside DeepSeek Harness to install this plugin; the full source repository is at https://github.com/Little-Star888/dsh-pelican. Restart the dsh process after installation for the plugin to take effect.

About this plugin

When an agent is thinking, the screen often just sits blank, leaving you unsure whether anything is actually happening. dsh-pelican fills that quiet gap with a little SVG pelican cycling lazily in the bottom-right corner, turning the wait into a small, pleasant visual moment.

On the host side the plugin maintains a three-state machine (idle / thinking / done) driven by global agent events such as inbox insertion and turn-stopping, and exposes a lightweight polling endpoint at /api/pelican/status. The client side animates the pelican in response: it lifts its head and whispers "I am here, just a moment~" while the agent is thinking, then pops a celebratory "All done! 🎉" bubble with a soft sound cue and an optional system notification once the task finishes. A built-in pelican_debug tool lets you inspect state sequences and event counts on the fly.

It is made for DSH Web users who want their agent workflow to feel a little warmer without introducing heavy logic. The plugin persists to disk and activates on restart, keeping zero footprint on live sessions.

Use Cases

  • Bottom-right pelican cycling animation signals an agent is actively thinking
  • Celebratory bubble, sound cue, and optional system notification on task completion
  • Inspect tri-state sequences and event counts via pelican_debug tool
  • Distinguish idle, thinking, and done at a glance across concurrent tasks

Best For

  • DSH Web users who prefer lightweight decoration without heavy logic
  • Indie developers who want warmth and personality in agent workflows
  • Teams that need visual task-status cues with zero runtime overhead