file-planning
Run the following command in DeepSeek Harness:
dsh plugin install JohnXu22786/file-planning
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install JohnXu22786/file-planning in your DeepSeek Harness terminal to install the plugin; the source repository is available at https://github.com/JohnXu22786/file-planning .
About this plugin
The hardest part of a multi-step task is rarely writing the plan. It is losing your place when a session dies, the context gets compressed, or you open a fresh window and have to guess where things left off. file-planning (Trail Map) writes the plan, execution state, field notes, and debrief entirely into a .trail/ directory in your workspace, so resuming just means reading the state and picking up exactly where you stopped.
Under the hood a small state machine governs each step through todo, in progress, and done, with the ability to cancel or re-open anytime. Steps can declare dependencies; unmet ones automatically flag downstream work as blocked, and circular dependencies are detected and refused. A template generator spins up a milestone skeleton for research, development, docs, or blank tasks with dependencies wired in one shot, and everything stays editable after creation. Every transition appends to an immutable event log, and the debrief command automatically summarizes milestones, steps, cancellations, and the timeline while preserving any hand-written retrospective sections you add. A final deterministic check confirms all steps are closed and the structure is sound before you call it done.
Built for DSH plugin-harness users running multi-step agent workflows who are tired of context loss. It ships with zero runtime dependencies, persists everything to plain files, and works standalone via a lightweight CLI or integrated through the dsh tools interface.
Use Cases
- A session dies mid-task; resume instantly by reading persisted state and picking up where you left off
- Research or dev work broken into dependency-linked steps with automatic blocking and completion tracking
- Wrap up with an auto-generated summary of milestones, cancellations, and timeline that stays auditable
Best For
- Developers running multi-step agent workflows inside the DSH plugin harness
- AI engineers who need task continuity across sessions and are tired of context loss
- Terminal users who want to manage multi-step plans with a zero-dependency lightweight CLI
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.