dsh-followup-todo
Run the following command in DeepSeek Harness:
dsh plugin install Tiaoma123/dsh-followup-todo
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install Tiaoma123/dsh-followup-todo in your terminal to install the plugin directly from https://github.com/Tiaoma123/dsh-followup-todo; it self-mounts via cordis.patch.yml and is ready to use immediately—no manual profile entry needed, and avoid mounting it twice to prevent duplicate prefix route errors.
About this plugin
The built-in todo_write is the agent's scratch pad for breaking down the current task, and it evaporates the moment the conversation ends. What you actually need is a different list: the conclusions you settled on, the key file paths, the hard constraints, recorded so that a brand-new session can pick up exactly where things left off without scrolling back through the old chat. dsh-followup-todo is that cross-session handoff note.
Each item carries two deliberately structured free-text fields—context (only confirmed facts, paths, boundaries) and inject (one-line goal, concrete next action, acceptance criteria, any ambiguity flagged as needs-confirmation). Together they compose a complete prompt; clicking the card drops it into the composer and the new session starts with full context in hand. The lifecycle supports P0–P3 priorities, mark-done, archive with full recall, and state lives in one plain-JSON file per workspace under ~/.dsh/, trivially backed up or cleared by hand.
On the browser side, a header badge button opens a floating panel for add / check / archive / delete across four filter tabs, and a collapsible card strip sits above the input dock, hiding itself when the list is empty. Only the web UI is implemented; headless profiles still get all four model tools (todo_add / todo_list / todo_done / todo_archive). Best suited to individual developers running long projects in DSH who hand context across conversations often. MIT-licensed, personal project, fork and modify freely.
Use Cases
- Push long projects across multiple conversations—each item carries its own context and next-step prompt so a new session starts without re-reading old chat
- Develop in parallel across workspaces with per-workspace file isolation and P0–P3 priority sorting
- Archive completed items with full recall, building a traceable log of decisions and tasks over time
Best For
- Solo developers running long projects in DSH who hand context across conversations often
- Users who need to turn informal discussion outcomes into structured handoff notes so a new session does not lose the thread
- Tinkerers who prefer plain-JSON data under their own control, backing up or clearing lists by hand
Related Plugins
Memory layer for coding agents that indexes local session history and auto-recalls relevant context before edits, commands, and failures, with no manual search needed.
Traceable, searchable cross-session memory for AI agents that turns conversation knowledge into a typed knowledge graph and recalls relevant subgraphs instead of replaying full history, natively integrated with DeepSeek Harness.
Gives DSH AI cross-session long-term memory, to-do and skill management, plus multi-session orchestration, external AI delegation, and an infinite canvas that grows with you.
dsh-mnemon is a three-tier, pluggable, Agent-driven memory system for DeepSeek Harness, combining Runtime memory, Project Documents, and replaceable Memory Spaces with nine long-term providers.