with one click
handoff
// Update a TPP with session progress and prepare it for the next session. Run when context is 80-90% full or when ending a session. Captures what was done, what was learned, and what comes next.
// Update a TPP with session progress and prepare it for the next session. Run when context is 80-90% full or when ending a session. Captures what was done, what was learned, and what comes next.
| name | handoff |
| description | Update a TPP with session progress and prepare it for the next session. Run when context is 80-90% full or when ending a session. Captures what was done, what was learned, and what comes next. |
| argument-hint | ["path/to/_todo/YYYYMMDD-name.md"] |
| disable-model-invocation | true |
| allowed-tools | Read, Glob, Grep, Edit, Write, Bash |
You are ending (or pausing) a session. Update the active TPP so the next session can pick up exactly where this one left off โ with no re-explanation needed.
If $ARGUMENTS is provided, use that path. Otherwise, find the most recently modified file in _todo/:
_todo/*.md, pick the most recently modifiedNote the existing structure so you preserve it. Do not rewrite sections that don't need updating.
Apply these edits, in order:
a. Current phase checkboxes
b. Tribal knowledge
c. Completed tasks
[x]d. Failed approaches (if any)
e. Remaining tasks
f. Next steps
## Next session starts here comment with 1โ3 bullet points naming the very first actions the next session should takeThe TPP must stay under 400 lines. After updating:
Report to the user:
Do not commit or push anything. The user will do that.