AI Agent Hub
Back to plugins
dsh-plugin-collapsible-steps preview

dsh-plugin-collapsible-steps

Client Updated 2026.08.25

Run the following command in DeepSeek Harness:

dsh plugin install mervyn-teo/dsh-plugin-collapsible-steps

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

Run dsh plugin install mervyn-teo/dsh-plugin-collapsible-steps in your terminal, then restart dsh web to activate the plugin; source code is available at https://github.com/mervyn-teo/dsh-plugin-collapsible-steps

About this plugin

A single turn in a DSH session can spawn a long chain of collapsible-card steps such as tool calls, workflow runs, compaction, context injection, model retries, and reasoning disclosures. Stacked end-to-end, dozens of these rows bury the final answer and make it painful to locate the one step that actually matters. dsh-plugin-collapsible-steps groups every run of consecutive steps into a single N-steps bracket row. One click collapses the entire run, hiding all but the most recent step and leaving no empty gaps; a second click expands it back. Because the plugin annotates the shipped conversation DOM instead of replacing any renderer, the original tool cards, reasoning sections, and markdown answers keep their native DSH appearance untouched. A global toggle called collapseByDefault in Settings lets you choose whether runs start collapsed or expanded by default. The plugin is tiny, consisting of just three lightweight script files with no additional configuration required. It is ideal for anyone running multi-step agent workflows in DSH who wants a cleaner, more scannable chat history without sacrificing the detail hidden inside each step card.

Screenshots

Use Cases

  • A long turn produces dozens of tool calls, retries, and compaction rows that bury the final answer
  • A team reviews agent execution logs and needs to jump to key steps instead of scrolling line by line
  • Repeatedly collapsing and expanding the same step group manually is slow and error-prone

Best For

  • Developers running multi-step agent workflows in DSH
  • Heavy users who find long step chains hard to read and want a cleaner chat history
  • Teams that need to quickly review and audit DSH session execution records