AI Agent Hub
Back to plugins
🤖

dsh-plugin-auxiliary-runtime

Model Inference Updated 2026.08.25

Run the following command in DeepSeek Harness:

dsh plugin install Hilbert-beinghappy/dsh-plugin-auxiliary-runtime

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

To install this plugin in DeepSeek Harness, run the command: dsh plugin install Hilbert-beinghappy/dsh-plugin-auxiliary-runtime, with the full source code available at https://github.com/Hilbert-beinghappy/dsh-plugin-auxiliary-runtime.

About this plugin

In AI toolchains, auxiliary model calls often present management challenges such as the inability to cancel mid-operation and incomplete usage tracking, leading to cost and security concerns. dsh-plugin-auxiliary-runtime addresses these issues by providing a comprehensive runtime solution for DeepSeek Harness.

The plugin's core capabilities include cancelable auxiliary model calls for safe interruption, a durable usage ledger that records token consumption for each call, and per-session policy limits to control concurrency and totals. It also offers three usage views—Official, Auxiliary, and Combined—for clear visibility into resource usage.

It is particularly suited for DeepSeek Harness users, especially developers extending model inference capabilities. If you're building plugins like Clarify or SeekTTY, this plugin simplifies auxiliary call management, enhancing overall efficiency and transparency.

Overall, dsh-plugin-auxiliary-runtime makes auxiliary model calls more controllable and transparent through reliable cancellation mechanisms and clear usage tracking, making it an essential tool in the Harness ecosystem.

Use Cases

  • Manage cancelable auxiliary model calls to avoid resource waste
  • Record and monitor token consumption per session for cost control
  • Provide a reliable auxiliary call runtime for plugins like Clarify and SeekTTY

Best For

  • Developers needing to extend DeepSeek Harness model inference capabilities
  • Engineers building integrated plugins seeking to simplify auxiliary call management
  • Teams looking to improve usage transparency and control in AI toolchains