planning-with-files

To maintain progress across a multi-step research project, use persistent markdown files as external working memory. Create a task_plan.md to track goals and phases, and save intermediate findings to notes.md to prevent context loss during iterative tool calls.

The best way to organize complex feature development workflows is using phased markdown checklists. Define goals in a task_plan.md, log status changes and errors iteratively, and produce a dedicated deliverable file for the final output.

Persistent file planning for bug fixes works by re-reading your task_plan.md before decisions and logging errors to support iterative agent loops. This ensures goals remain accessible across many tool calls without relying on ephemeral context.

Yes, you can use markdown planning files for iterative agent loops and automation tasks. The workflow requires creating and updating task_plan.md, writing findings to notes.md, and logging status changes to support repeated tool calls and checkpoints.

No specific dependencies are required to manage structured research notes with markdown. Only need to create task_plan.md and notes.md files to start organizing complex tasks into phased checklists and intermediate findings.

Avoid using persistent markdown files for workflow planning when tasks are simple enough to fit within ephemeral context. It is designed for complex, multi-step tasks that benefit from phased checklists and repeated tool calls.