AI Agent Hub
Back to plugins
⚙️

cleverer-dsh

Workflow Updated 2026.08.25

Run the following command in DeepSeek Harness:

dsh plugin install Classicoke/cleverer-dsh

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

To install this plugin in DeepSeek Harness, run `dsh plugin install Classicoke/cleverer-dsh` or visit the full source repository at https://github.com/Classicoke/cleverer-dsh for script-based installation options.

About this plugin

While DeepSeek Harness (DSH) is highly capable, its default behavior can sometimes lack smart execution—stubbornly retrying failed commands, ignoring task planning, and failing to utilize installed skills. cleverer-dsh is a comprehensive plugin suite designed to fix these pain points by enforcing strict execution discipline. It teaches the Agent to automatically intercept failures, force reflection, and update plans, effectively eliminating meaningless retry loops.

At its core, the suite combines eight discipline plugins with a unified hub to orchestrate Agent behavior, alongside practical tools like instant multi-directory file lookup and a 9-point environment health check. It also features six built-in skills that auto-load on demand. Real-world tests show that this structured approach significantly reduces LLM and tool calls, cutting down token consumption and total task completion time.

If you are tired of your Agent getting stuck on the same errors, or if you want DSH to act like a seasoned engineer who plans before executing and learns from mistakes, this zero-dependency suite is the perfect upgrade. It is ideal for DSH users seeking higher efficiency, lower costs, and a highly disciplined agent workflow.

Use Cases

  • Automatically intercepting errors and forcing reflection to prevent the agent from getting stuck in infinite retry loops.
  • Prompting task planning before execution and dynamically updating to-do lists during complex workflows.
  • Quickly scanning multiple directories and checking environment health to eliminate blind guessing and inefficient exploration.

Best For

  • DSH users looking to reduce token consumption and improve task completion efficiency.
  • Developers tired of stubborn agent retries who need standardized execution discipline.
  • Plugin users seeking out-of-the-box, zero-dependency solutions with high test coverage.