dsh-acp
Run the following command in DeepSeek Harness:
dsh plugin install cnctem/dsh-acp
Paste the following prompt into your AI chat to install this plugin:
To install the dsh-acp plugin in DeepSeek Harness, run the command `dsh plugin install cnctem/dsh-acp` or clone the source from the GitHub repository at https://github.com/cnctem/dsh-acp and build it.
About this plugin
In the realm of modern AI-assisted development, seamlessly integrating intelligent agents into IDE workflows is crucial, yet standard tools often lack deep interactivity. The dsh-acp plugin addresses this gap by acting as a bridge between DeepSeek Harness and IDEs like Zed, enabling agent-driven local sessions via JSON-RPC 2.0. This allows developers to activate, control, and observe AI agents directly within their editing environment, streamlining the fusion of AI and coding tasks.
At its core, dsh-acp transforms complex agent behaviors into a rich, IDE-native experience. It supports real-time token and thinking streaming, tool call cards with location details, structured diffs for edits, session history management, and preset agent configurations. Notably, it extends capabilities to include image understanding—allowing agents to process screenshots—and slash commands with form-based elicitation, enabling interactive planning, debugging, and multimodal assistance that feel intuitive and powerful.
This plugin is ideal for developers already leveraging DeepSeek Harness as their agent backend and seeking an enhanced experience in Zed or other ACP-compatible IDEs. Whether for daily coding, complex project debugging, or scenarios requiring multimodal interaction, dsh-acp brings the agent's capabilities into a smooth, editor-integrated workflow, boosting productivity and natural collaboration with AI.
Screenshots
Use Cases
- Drive AI agents in Zed editor for real-time code completion and debugging.
- Manage agent sessions via JSON-RPC protocol to track task progress and history.
- Leverage image understanding to process screenshots for enhanced multimodal interaction.
Best For
- Developers using DeepSeek Harness as their agent backend.
- Users who prefer Zed or other ACP-compatible IDEs.
- Engineers requiring AI-assisted coding, debugging, and project planning.
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.