Keeping AI coding context across sessions with HANDOFF.md
This Reddit post describes a problem in long AI development sessions: the AI loses context when a new session starts. The author presents a workflow with capture and handoff sub-skills that write discoveries and session state into a HANDOFF.md file. They say hooks can automatically add that file to new sessions so the AI can continue from the previous work.
Key points
- The post focuses on context loss in long AI development sessions.
- HANDOFF.md stores discoveries, current state, and handoff notes.
- The workflow uses capture and handoff sub-skills.
- Hooks are used to inject HANDOFF.md into new sessions automatically.
- The post labels the workflow as advanced.
Quick term guide
- sessions
- Separate work threads or task runs inside a tool.
- workflow
- A repeatable set of steps for getting a task done.
- sub-skills
- Smaller reusable abilities inside a larger AI workflow.
- HANDOFF.md
- A markdown file used to record the current work state for the next session.
- Solo makers
- People who build and launch their own products or services entirely on their own.
- pain point
- A specific problem or frustration that people experience repeatedly and want solved
- background
- Running out of sight while the main app or screen stays focused on something else.
- session reset
- Starting the current working session over.