dsh-workstate
Run the following command in DeepSeek Harness:
dsh plugin install luoyuejun9/dsh-workstate
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install luoyuejun9/dsh-workstate inside DeepSeek Harness to install this plugin; source code is at https://github.com/luoyuejun9/dsh-workstate .
About this plugin
In multi-agent coding workflows, traditional handoffs rely on prose descriptions of what was done. The next agent has no machine-readable view of progress, decisions, failures, or verification results, and no way to confirm the workspace is unchanged. dsh-workstate bridges this gap by packaging the current task state, key decisions, verification outcomes, changed paths, and a Git integrity fingerprint into a structured state packet, turning handoff from reading a memo into reading data.
The core capability revolves around four actions: checkpoint pauses and preserves the current work; transfer creates a shareable export for another agent or developer; resume injects an existing packet and validates workspace consistency before proceeding; and diff compares two checkpoints by stable work-item IDs. The plugin never mutates Git state, with no checkout, reset, commit, or push, keeping safety by design.
It is well suited for teams and individual developers running multi-agent coding flows in DeepSeek Harness. If you need different agents to pick up the same code task in sequence, or to hand work reliably between agents and humans, dsh-workstate offers a lightweight, local-first, protocol-driven alternative.
Use Cases
- Handing off a coding task between AI agents mid-stream
- Resuming a paused agent session with validated workspace state
- Diffing two checkpoints to track work progress over time
- Reliable task handoff between an agent and a human developer
Best For
- Developers orchestrating multi-agent coding flows with DeepSeek Harness
- Teams that need reliable task handoff between agents and humans
- Engineering teams that prioritize Git safety in their tooling
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.