long-horizon-runtime
Run the following command in DeepSeek Harness:
dsh plugin install ynymhrb/long-horizon-runtime
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install ynymhrb/long-horizon-runtime in your DeepSeek Harness terminal to install this long-task plugin from https://github.com/ynymhrb/long-horizon-runtime.
About this plugin
Research, development, debugging, and documentation work rarely wraps up in a single chat exchange. Pushing these long tasks through a plain conversation means progress lives buried in message history: pause once and you have to dig back in, pivot the goal and completed work is at risk of being lost. Long Horizon Runtime turns a long task into a structured workflow: it first presents a step-by-step plan for your confirmation, then executes; every stage status and dependency is visible in the Task Area without scrolling through chat.
The core capabilities centre on three things: confirmable plans, real-time progress visibility, and resumable checkpoints. If you need to change direction mid-run, completed steps are preserved and the plugin generates a fresh follow-up plan around your new goal rather than starting over. If the conversation is paused or interrupted, you simply return to the same task and pick up where it left off, no need to reconstruct context yourself.
It is well suited to multi-step, steadily advancing work: writing research reports, debugging complex systems, organising large document sets, or completing phased development tasks. No mode switch is required; just describe what you need done and the plugin shapes it into a trackable long task, letting you work in a review-the-plan-first-then-execute rhythm and see a long job through to completion.
Screenshots
Use Cases
- Advancing multi-stage research, development or debugging work
- Resuming unfinished steps after a conversation is interrupted
- Changing direction mid-task while preserving completed work
Best For
- Developers driving multi-step work forward iteratively
- Knowledge workers handling research or document curation
- Users frequently disrupted by lost progress in long conversations
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.