# TODOs

- [x] Complete the public documentation set under `docs/`, covering getting started, authoring, commands/extraction configuration, interaction/accessibility, results, troubleshooting, security/privacy, architecture, development/testing, and releasing/deployment.
- [x] Make the extension's extraction model configurable instead of fixed to one model:
  - On first extraction use, offer a concise choice between using the session model with low extraction thinking and choosing a dedicated provider, model, and thinking level, then store the choice globally.
  - Provide a searchable picker for Pi's authenticated models and initially highlight/recommend low thinking for speed and lower cost.
  - If a custom model fails, fall back to the session model and session thinking setting.
- [x] Use display name **Pi Rich Questions**, project/repository and npm package `pi-rich-questions`, extension directory `extensions/pi-rich-questions`, slash command `/rich-questions`, and tool API `ask_rich_questions`.
- [x] Upgrade version `0.4.0` to Pi's `0.84.4` model runtime and `0.85.x` fullscreen interaction APIs, including constrained sampling, structured extraction, sequential tool execution, refresh cancellation, configured keybindings, mouse support, compatibility tests, and release documentation.