菜单
Reproduce and isolate auto-aim detection, tracking, solving, aiming, and planning issues from recorded video in this repository. Use for false positives, missed armors, tracker state changes, aim jitter, fire decisions, or latency regressions.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/HNUST-Dust/Dust_infantry_vision --skill replay-auto-aim命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Audit YAML, model paths, OpenVINO device selection, camera and serial settings, and camera/gimbal calibration values for this repository. Use when parameters appear ignored, deployment differs between machines, or pose and model behavior are suspicious.
Review the changes since a fixed point (commit, branch, tag, or merge-base) along two axes — Standards (does the code follow this repo's documented coding standards?) and Spec (does the code match what the originating issue/spec asked for?). Runs both reviews in parallel sub-agents and reports them side by side. Use when the user wants to review a branch, a PR, work-in-progress changes, or asks to "review since X".
Shared vocabulary for designing deep modules. Use when the user wants to design or improve a module's interface, find deepening opportunities, decide where a seam goes, make code more testable or AI-navigable, or when another skill needs the deep-module vocabulary.
基于 SOC 职业分类
正在显示 SKILL.md
| name | replay-auto-aim |
| description | Reproduce and isolate auto-aim detection, tracking, solving, aiming, and planning issues from recorded video in this repository. Use for false positives, missed armors, tracker state changes, aim jitter, fire decisions, or latency regressions. |
Use recorded inputs before using a camera or gimbal. The goal is a tight, repeatable loop that can go red on the reported symptom.
assets/demo/demo.avi and assets/demo/demo.txt, select a small frame range, and run from the repository root.detector_video_test for YOLO versus traditional detection and corner geometry;auto_aim_test for YOLO, Solver, Tracker, and Aimer together;planner_test_offline for trajectory and fire planning without camera or gimbal.device: CPU; never overwrite the robot config during diagnosis.auto_aim_test and detector_video_test use OpenCV windows and require a display; report headless limitations explicitly.Plotter sends JSON to UDP 127.0.0.1:9870; capture or disable that observer only when it is part of the repro.$diagnosing-bugs for difficult reproductions and $tdd for the retained regression test.Read replay-matrix.md for command syntax and pipeline facts.