Directories
¶
| Path | Synopsis |
|---|---|
|
cmd/attack_record is a MACRO RECORDER + REPLAYER for ClashGO.
|
cmd/attack_record is a MACRO RECORDER + REPLAYER for ClashGO. |
|
Command attack_report answers the two questions a finished attack actually raises, from the run's own log:
|
Command attack_report answers the two questions a finished attack actually raises, from the run's own log: |
|
Command attack_verify is the 720p review harness: it proves the attack plan is correct BEFORE a battle is spent on it, by rendering exactly what the bot would tap onto a live (or saved) 720p battle screenshot.
|
Command attack_verify is the 720p review harness: it proves the attack plan is correct BEFORE a battle is spent on it, by rendering exactly what the bot would tap onto a live (or saved) 720p battle screenshot. |
|
cmd/capture_template is a manual template-recapture tool.
|
cmd/capture_template is a manual template-recapture tool. |
|
Command classify_probe classifies a screenshot (default: live adb capture) with the repo's own classifier and prints the score for every state rule plus a few key pixel reads, so a stuck bot can be diagnosed from what the vision layer actually sees.
|
Command classify_probe classifies a screenshot (default: live adb capture) with the repo's own classifier and prints the score for every state rule plus a few key pixel reads, so a stuck bot can be diagnosed from what the vision layer actually sees. |
|
cmd/design_attack is the one-stop tool for authoring + verifying the per-unit `formula.json` consumed by the deploy pipeline.
|
cmd/design_attack is the one-stop tool for authoring + verifying the per-unit `formula.json` consumed by the deploy pipeline. |
|
Command pick_coords is the deploy-point picker: it shows the live emulator frame in the Preview tab and lets you click the exact points the bot should use, per side, then writes them into precision_config.json.
|
Command pick_coords is the deploy-point picker: it shows the live emulator frame in the Preview tab and lets you click the exact points the bot should use, per side, then writes them into precision_config.json. |
|
Command release_manifest emits a `latest.json` file for the just-built release zip.
|
Command release_manifest emits a `latest.json` file for the just-built release zip. |
|
Command resprobe measures how the CoC HUD scales between two framebuffer geometries.
|
Command resprobe measures how the CoC HUD scales between two framebuffer geometries. |
|
Command result_probe is a diagnostic harness that runs the bot's own battle-result OCR pipeline against a saved screenshot (defaults to last_battle_result.png in the config dir).
|
Command result_probe is a diagnostic harness that runs the bot's own battle-result OCR pipeline against a saved screenshot (defaults to last_battle_result.png in the config dir). |
|
Command see is the debugging agent's eyes on the emulator.
|
Command see is the debugging agent's eyes on the emulator. |
|
Command swipe_probe is a standalone live-verification harness for the human-gesture input layer (Phase 1 of the autonomous hardening loop): it connects to the emulator like the bot does and exercises the sendevent bezier-swipe primitives against the live device.
|
Command swipe_probe is a standalone live-verification harness for the human-gesture input layer (Phase 1 of the autonomous hardening loop): it connects to the emulator like the bot does and exercises the sendevent bezier-swipe primitives against the live device. |
|
cmd/test_wall_upgrade is a manual verification tool for the wall-upgrade system.
|
cmd/test_wall_upgrade is a manual verification tool for the wall-upgrade system. |
|
Command tplprobe locates one stored template on a frame across a scale sweep and prints the ranked result.
|
Command tplprobe locates one stored template on a frame across a scale sweep and prints the ranked result. |
|
Command wmresize changes a running device's display geometry.
|
Command wmresize changes a running device's display geometry. |
Click to show internal directories.
Click to hide internal directories.