Coordinate long-running tasks with persistent markdown planning files.
npx skills add https://github.com/aleph23/Natasha --skill planning-with-files-aleph23
Please help me install this Agent Skill. Skill: planning-with-files Source: https://github.com/aleph23/Natasha/tree/main/skills/planning-with-files Command: npx skills add https://github.com/aleph23/Natasha --skill planning-with-files-aleph23
planning-with-files solves the challenge of managing long-running, multi-step tasks by using persistent markdown files as a "working memory on disk" to maintain plans, findings, and progress across sessions.
Initialize a new planning session by creating task_plan.md, findings.md, and progress.md in your project folder.
High-intent search queries and answers about installing and using this skill.
Automatic session recovery for long-running tasks works by maintaining persistent markdown files as working memory on disk. By continuously updating task_plan.md, findings.md, and progress.md through read and write hooks, the system keeps plan state synchronized with tool activity and allows seamless continuity after resets.
You can plan complex multi-step tasks by initializing task_plan.md, findings.md, and progress.md directly in your project folder. This establishes a structured planning session that enforces plan-driven execution, coordinates tool calls, and automatically logs progress as work proceeds.
Yes, markdown files support session recovery after a clear by acting as a persistent working memory on disk. Files like task_plan.md, findings.md, and progress.md store your goals and session logs, allowing you to automatically resume multi-phase projects without losing context.
File-based planning is ideal for complex audits and iterative decision making because it maintains task_plan.md, findings.md, and progress.md to document goals and research. This enforces plan-driven execution and keeps plan state in sync with tool activity throughout multi-phase projects.
You need persistent file-based memory for task planning when managing long-running, multi-step tasks that require multiple tool calls and iterative decisions. It enforces plan-driven execution and enables automatic session recovery, ensuring continuity across sessions by maintaining plans, findings, and progress.
H2OSLabsMaintain task_plan.md, findings.md, and progress.md files for multi-step tasks.
Shawn-Guo-CNCreates and manages task_plan.md, findings.md, and progress.md files for persistent task planning.
minicooheiInitialize and update task_plan.md, findings.md, and progress.md files for multi-phase projects.
barkleesandersCreate and track file-based todos for code review feedback and technical debt.
