Foreword

In AI agent development and operations practices, measuring system reliability is the foundation for building stable services. DeepSeek Harness (DSH) adopts a plugin-based architecture, allowing developers to implement specific functions by extending skills. The satan9394/dsh-slo plugin is designed for this scenario, bringing core concepts from SRE (Site Reliability Engineering) into the DSH ecosystem.

Plugin Overview

This plugin belongs to the “Workflow” category and is maintained by the user satan9394.

  • Plugin Name: satan9394/dsh-slo
  • Core Purpose: DSH skill for measuring SLI/SLO/error budgets and supporting SRE practices
  • Design Reference: Inspired by the wshobson/agents project

Core Capabilities

Based on verified facts, this plugin aims to help users implement SRE practices in DSH environments:
* Provides measurement capabilities related to SLI (Service-Level Indicators).
* Provides setup and tracking capabilities for SLO (Service-Level Objectives).
* Provides capabilities for calculating and managing Error Budgets.

Installation and Usage

Because specific installation commands are not provided, please refer to the following official channels to obtain the plugin source code and documentation:

Catalog: https://www.skillhub.cn/plugins/satan9394/dsh-slo
GitHub:  https://github.com/satan9394/dsh-slo

When using it, make sure the plugin is properly loaded into the DSH process, and configure the relevant SRE parameters according to its documentation.

Use Cases and Cautions

  • Intended Audience: Developers and operations personnel who need to implement SRE practices in AI agents or automated workflows.
  • Permission Check: The plugin runs with the permissions of the current DSH process. Before installing or using it, review its source code to ensure compliance with security standards.

Conclusion

The satan9394/dsh-slo plugin provides a tool for quantifying reliability objectives in the DSH ecosystem, helping integrate SRE principles into AI agent development workflows. For more details and the latest updates, visit its catalog page or GitHub repository.