AI Agent Hub
Back to plugins
🤖

dsh-superpowers-zh

Model Inference Updated 2026.08.21

Run the following command in DeepSeek Harness:

dsh plugin install fatherplus/dsh-superpowers-zh

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

Run dsh plugin install fatherplus/dsh-superpowers-zh in your terminal, then restart the DSH Web service to activate the plugin; source lives at https://github.com/fatherplus/dsh-superpowers-zh .

About this plugin

When running model inference with DeepSeek Harness, engineering methodology often degrades into ad-hoc prompt copying and one-off Markdown files scattered across a repo. dsh-superpowers-zh bundles the eight core engineering skills from jnMetaCode/superpowers-zh into a single lightweight adapter, so you can load them all at once without manually porting each file into your Harness configuration.

The eight skills form a complete engineering loop: brainstorming, writing plans, executing plans, systematic debugging, test-driven development, requesting code reviews, receiving code reviews, and verification before completion. The plugin is a thin adapter that registers these upstream skill files with DSH and deliberately stays out of the way—it ships no Pi-specific extensions, workflow runners, credential management, or local configuration logic, keeping the install footprint minimal and the update surface small.

It is a good fit for developers who already build prompt chains or inference pipelines on DSH and want structured engineering practices baked into their workflow, as well as teams that prefer a one-command way to stay in sync with upstream skill updates rather than hand-maintaining a growing pile of Markdown files.

Use Cases

  • Loading structured engineering loops into DSH inference pipelines
  • One-command sync of upstream superpowers-zh skill updates
  • Embedding TDD, debugging, and code review into the workflow

Best For

  • Developers building inference pipelines on DSH
  • Teams wanting to bake engineering practices into prompt chains
  • Engineers preferring one-command skill updates over manual file maintenance