AI Agent Hub
Back to plugins
🤖

dsh-percentile

Model Inference Updated 2026.08.21

Run the following command in DeepSeek Harness:

dsh plugin install uckkk/dsh-percentile

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

Run dsh plugin install uckkk/dsh-percentile in DeepSeek Harness to install this plugin; the source code is available at https://github.com/uckkk/dsh-percentile.

About this plugin

In model inference pipelines, percentile is a critical statistical indicator for measuring output distribution, yet many solutions require calling external services or making network requests, adding both latency and availability risk. dsh-percentile addresses this pain point by packaging percentile calculation as a plug-and-play DeepSeek Harness plugin.
The plugin is implemented in pure Node.js and produces zero network dependencies during execution, completing all computations locally. This makes it a natural fit for latency-sensitive inference workflows or offline environments where deploying additional network services is impractical.
If you are building a model inference chain and need to introduce percentile statistics for output scoring or quality monitoring while keeping network overhead at zero under the permissive MIT license, dsh-percentile is a lightweight option worth adding to your toolkit.

Use Cases

  • Computing percentile metrics in model output quality monitoring
  • Zero-network-overhead statistical analysis in offline inference pipelines
  • Localized statistical computation in latency-sensitive inference workflows

Best For

  • Engineers building model inference pipelines
  • Development teams requiring localized statistical computation
  • Offline deployment scenarios demanding zero network dependencies