An AI agent built to do Ralph loops - plan mode for planning and ralph mode for implementing.

docs: add more TODOs

+3 -2
+3 -2
TODO.md
··· 6 6 - [x] Spinner animation 7 7 - [x] Help overlay 8 8 - [x] Keyboard navigation in Kanban 9 + - [x] Thinking indication when running - no idea what model is doing 9 10 - [ ] Resumable conversations/sessions 10 11 - [ ] Wrapping/dynamically sized input for planning chats 11 - - [ ] replace tool calls with tool results in planning chat view 12 + - [ ] Replace tool calls with tool results in planning chat view 12 13 - [ ] Respond better to 429 requests 13 14 - [ ] New critique agent that will ruthlessly evaluate specs against plans 14 15 - [ ] Configurable prompts for all agents 15 16 - [ ] Utilize AGENTS.md if it exists 16 17 - [ ] Focus on token minimization and cost control - maybe default to Haiku for implementation? Opus for critique, Sonnet for planning? 17 - - [ ] Thinking indication when running - no idea what model is doing 18 + - [ ] Configurable editing mode for planning, should default to insert