Todo

A working memory todo tool for conversation-scoped task planning.

This tool provides a lightweight todo list that exists within the current conversation context, complementing the existing persistent task management system in gptme-agent-template.

Key principles:

• Working Memory Layer: Ephemeral todos for current conversation context

• Complements Persistent Tasks: Works alongside existing task files without conflicts

• Simple State Model: pending, in_progress, completed, paused

• Conversation Scoped: Resets between conversations, doesn’t persist to disk

• Auto-replay: Automatically restores todo state when resuming conversations

gptme.tools.todo.get_incomplete_todos_summary() → str

Get a summary of incomplete todos for continuation prompts.

Returns:

A formatted string listing incomplete todos, or empty string if none.

gptme.tools.todo.has_incomplete_todos() → bool

Check if there are any incomplete todos in working memory.

Used by auto_reply_hook to determine if the agent should continue working instead of being asked about completion.

Returns:

True if there are any pending or in_progress todos.