dsh-agent-bridge
Run the following command in DeepSeek Harness:
dsh plugin install CAI-MH/dsh-agent-bridge
Paste the following prompt into your AI chat to install this plugin:
To install this plugin in DeepSeek Harness, run dsh plugin install CAI-MH/dsh-agent-bridge. Source code is available at https://github.com/CAI-MH/dsh-agent-bridge
About this plugin
Algorithm teams often discuss approaches in a Feishu group, but verifying an idea still means switching back to the local environment: opening a session manually, writing a persona, attaching a knowledge base, and pasting the result back. dsh-agent-bridge collapses that loop into a single step. A bot running as an independent daemon delivers the task over localhost; the plugin opens a real DSH session in the target workspace, injects the specified persona, skills, and knowledge base, executes with DSH's existing LLMs and all installed plugins, and posts the final text back to the Feishu group. The entire surface listens only on 127.0.0.1 and requires a shared secret, so nothing is exposed to the wider network.
Workspace routing is straightforward: if the task text mentions a DSH workspace title or path (whitespace-insensitive), the bridge dispatches the task there. If no workspace can be identified, it returns a clear error asking the user to name one, preventing execution in the wrong environment. A brand-new task spawns a fresh session; follow-up questions on the same result reuse the existing session, preserving context.
It is built for small teams or individual researchers who already rely on DSH Desktop for daily development and use Feishu groups for collaboration. Mention the task in the group, and the result lands back in the same thread within minutes—no copy-paste, no separate remote server, and no data leaving the local machine.
Use Cases
- Describe an algorithmic question in a Feishu group and receive a detailed local-LLM answer minutes later
- Follow-up questions automatically reuse the previous session with full context preserved
- Mention a workspace title or path in the task text and the bridge routes execution to the correct environment
Best For
- Algorithm engineers who use DSH Desktop for daily local development
- Small tech teams that discuss in Feishu groups but verify in a local environment
- Researchers who want group discussions to drive local execution without repeated copy-paste
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.