AI Agent Hub
Back to plugins
dsh-skill-manager preview

dsh-skill-manager

Client Updated 2026.08.24

Run the following command in DeepSeek Harness:

dsh plugin install LongSir0419/dsh-skill-manager

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

Install the plugin by running the install command in your DeepSeek Harness terminal; the source code is available at https://github.com/LongSir0419/dsh-skill-manager

About this plugin

User-level Skills in DSH live as directories under $DSH_HOME/skills, each paired with a SKILL.md file. Tinkering with descriptions, flipping frontmatter flags, renaming directories, or adding and removing entries all required opening a text editor and editing lines by hand — error-prone and hard to survey at a glance.

dsh-skill-manager moves that entire workflow into the Web settings panel. Open Settings → Skill Manager and every Skill shows up in a clean list. Click a row to expand the full SKILL.md content and edit it inline; flip the toggle on the right to enable or disable a Skill instantly — under the hood it rewrites the disable-model-invocation field in the frontmatter, and DSH's filesystem watcher immediately excludes that Skill from the model catalog without a restart. Creating a new Skill, renaming it (directory name and frontmatter stay in sync), and deleting it (with a confirmation step) are all one-step operations.

If you maintain several user-level Skills day to day or frequently toggle their enabled state to switch model behaviour, this plugin lets you manage the whole Skill collection through a visual interface instead of editing markdown line by line in a text editor.

Screenshots

Use Cases

  • View, edit, or create SKILL.md content from the Web settings panel
  • Toggle a Skill on or off instantly without restarting DSH
  • Rename or delete a user-level Skill with directory and frontmatter auto-synced

Best For

  • DSH users who maintain multiple user-level Skills day to day
  • Operators who prefer a visual UI over manual file editing
  • Developers who frequently toggle Skill states to test model behaviour