AI Agent Hub
Back to plugins
⚙️

dsh-session-bridge

Workflow Updated 2026.08.15

Run the following command in DeepSeek Harness:

dsh plugin install FuWenLianggit/dsh-session-bridge

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

Run dsh plugin install FuWenLianggit/dsh-session-bridge in the DeepSeek Harness terminal to install this plugin from the GitHub repository at https://github.com/FuWenLianggit/dsh-session-bridge, then restart dsh web and hard-refresh your browser to activate it.

About this plugin

DeepSeek Harness sessions lack a native mechanism for cross-session messaging or orchestration. Agents cannot spawn peer sessions, dispatch prompts to them, or inspect their run state without external tooling. dsh-session-bridge closes that gap by wrapping the harness's built-in session and agent services into model-facing tools, turning inter-session collaboration into a first-class capability.

Three tools are exposed: session_create spins up an ordinary session in the caller's workspace and inherits the current composition configuration; session_send delivers a prompt to any ordinary session by id, offering both queue (next-turn FIFO) and interrupt (next-step boundary) delivery modes with automatic cold-session resume; session_list enumerates ordinary sessions along with their running state. All tools register globally in the web profile, independent of any agent preset, so every session can invoke them directly.

The plugin suits agent workflows that require task fan-out, parallel orchestration, or cross-session status aggregation. A lead agent can delegate sub-tasks to multiple peer sessions and collect results without manual intervention. Every cross-session message carries a plugin provenance tag for traceability and auditing, making it easy to attribute who initiated what action.

Use Cases

  • A lead agent delegates sub-tasks in parallel to multiple peer sessions and collects results
  • Dispatching prompts across sessions with FIFO queue or interrupt delivery modes
  • Enumerating and monitoring the running state of all ordinary sessions in a workspace

Best For

  • Agent workflow developers who need parallel multi-session orchestration
  • Harness users who want task fan-out with result aggregation
  • Platform engineers who require cross-session message audit trails