dsh-plugin-longgraph
Run the following command in DeepSeek Harness:
dsh plugin install levi-qiao/dsh-plugin-longgraph
Paste the following prompt into your AI chat to install this plugin:
In DeepSeek Harness, install this plugin by running 'dsh plugin install levi-qiao/dsh-plugin-longgraph', with the full source code available at https://github.com/levi-qiao/dsh-plugin-longgraph.
About this plugin
When developing complex, long-horizon agent workflows, developers often face challenges such as disorganized prompt management and difficult context maintenance. dsh-plugin-longgraph, a community plugin for DeepSeek Harness, addresses these issues by providing a structured prompt library to streamline the orchestration of long-term agent tasks.
The plugin's core capabilities revolve around registering three key skills: longgraph, loop-graph, and loop-converge. These skills support the entire process from fit checks and user interviews to compiling configurations and launching workflows, ensuring that agents maintain clear context and efficient instruction delivery during extended operations. It does not implement an agent runtime but instead focuses on skill-level extensions, allowing developers to concentrate on workflow design.
This plugin is ideal for developers using DeepSeek Harness to build complex agent systems, particularly researchers or engineers who need to manage multi-step, long-duration tasks. If you are seeking a reliable way to organize and optimize agent workflows, this plugin provides the necessary tools and flexibility.
Use Cases
- Used when developing complex, multi-step agent tasks.
- Optimizing prompt management and context maintenance in agent workflows.
- Integrating skills in DeepSeek Harness to support long-duration tasks.
Best For
- Developers building long-horizon agent systems.
- Researchers focused on workflow efficiency and reliability.
- Engineers or technical experts using DeepSeek Harness.
Related Plugins
A method pack that makes AI coding agents plan against your real baseline, prove completion with fresh evidence, and reduce reworks and unsafe changes.
Turns the DeepSeek Harness session into a captain that builds a durable sub-agent team, splits goals into dependency-aware tasks, and coordinates work via direct messages and a live Web UI.
Gives coding agents design judgment, letting Claude Code, Cursor, and 70+ agents generate and iterate high-quality UI, presentations, and graphics right from the terminal.
Run the Pi ecosystem's plugins on DeepSeek Harness, unmodified, via a compatibility layer that implements Pi's public extension ABI on DSH's native services.