ichikas-auto-assistant

Add or update an Ichika Auto Assistant task by wiring everything around the task implementation itself. Use when Codex needs to create a new task placeholder, register a task, add scheduler enable switches, extend config models and default JSON files, expose task settings in the desktop GUI, or verify that a new task is reachable from CLI and GUI without implementing the business logic yet.

Integrate reusable image template resources into the IAA repository through the standard resources-to-R.py pipeline. Use when Codex needs to turn a captured image patch into a formal template under `resources/`, add or update a `*.png.json` resource definition, run `tools/make_resources.py`, verify the generated prefab appears in `iaa/tasks/R.py`, or replace ad-hoc image path loading with generated `R.*` references.

Debug live game UI automation through kotonebot and the emulator. Use when Codex needs to connect to MuMu, initialize kotonebot context, control the device, capture screenshots, save runtime artifacts, write one-off probe scripts, or investigate why a recognition/click/switch flow behaves incorrectly on a real page.

Debug and refine kotonebot UI recognition logic. Use when Codex needs to analyze ROI choice, HSV thresholds, mask quality, morphology, contours, candidate scoring, cross-frame stability, or the visual semantics of what should be detected on game UI screenshots.