重要前提
安装AI Skills的关键前提是:必须科学上网,且开启TUN模式,这一点至关重要,直接决定安装能否顺利完成,在此郑重提醒三遍:科学上网,科学上网,科学上网。查看完整安装教程 →
npx skills add https://github.com/marswangyang/workout-tracker --skill workout-tracker一个使用本地 SQLite 数据库记录和追踪健身活动的技能。
当用户说"我做了 5 组 60 公斤的卧推,每组 8 次"时,解析详情并运行:
cd skills/workout-tracker/scripts && uv run log.py \
--exercise "Bench Press" \
--weight 60 \
--unit kg \
--sets 5 \
--reps 8 \
--body_part "Chest" \
--rest_time 90 # 可选:用户提供或留空以自动填充
--notes "Optional notes"
参数:
--exercise (必需):动作名称。--body_part:(新增) 目标肌群 (例如:胸部、背部、腿部)。如果未提供,请根据锻炼项目推断 (例如:卧推 -> 胸部)。--weight (必需):重量值。广告位招租
在这里展示您的产品或服务
触达数万 AI 开发者,精准高效
--unit:'kg' 或 'lb' (默认为 kg)。--sets:组数。--reps:每组次数。--rest_time:秒数。如果用户未指定,请省略此参数。 脚本将从历史记录中自动填充。当用户询问"我的深蹲进展如何?"或"显示最近的训练"时:
cd skills/workout-tracker/scripts && uv run render.py --exercise "Squat"
然后发送生成的图片: message(action="send", filePath="/tmp/workout_report.png", message="这是你的报告:")
如果用户上传了一张图片 (例如,一台器械):
log.py 脚本。body_part 从锻炼名称 (例如:"Bench Press" -> "Chest"),无需询问,除非有歧义。render.py 为报告生成图片。figsize 宽度 ~6)。0。数据存储在 skills/workout-tracker/workout.db (SQLite) 中。模式包括 date、exercise、weight、unit、sets、reps、rest_time、notes。
此技能使用 uv 进行 Python 包管理。
确保已安装 uv:brew install uv
同步依赖项:
cd skills/workout-tracker/scripts uv sync
SQLite 数据库 (workout.db) 使用以下模式:
id:整数 (主键)date:ISO8601 字符串exercise:字符串body_part:字符串 (目标肌群)weight:浮点数unit:字符串 (默认 "kg")sets:整数reps:整数rest_time:整数 (秒数,可选)notes:字符串 (可选)每周安装次数
50
仓库
首次出现
2026年1月30日
安全审计
安装于
gemini-cli47
codex47
opencode47
github-copilot46
cursor46
cline45
A skill to log and track fitness activities using a local SQLite database.
When the user says "I did 5 sets of Bench Press at 60kg for 8 reps", parse the details and run:
cd skills/workout-tracker/scripts && uv run log.py \
--exercise "Bench Press" \
--weight 60 \
--unit kg \
--sets 5 \
--reps 8 \
--body_part "Chest" \
--rest_time 90 # Optional: User provided or leave empty to auto-fill
--notes "Optional notes"
Parameters:
--exercise (Required): Name of the movement.--body_part: (New) Target muscle group (e.g. Chest, Back, Legs). INFER THIS from the exercise if not provided (e.g. Bench -> Chest).--weight (Required): Weight value.--unit: 'kg' or 'lb' (default kg).--sets: Number of sets.--reps: Reps per set.--rest_time: Seconds. OMIT THIS if the user didn't specify. The script will auto-fill from history.When the user asks "How is my Squat progress?" or "Show last workouts":
cd skills/workout-tracker/scripts && uv run render.py --exercise "Squat"
Then send the generated image: message(action="send", filePath="/tmp/workout_report.png", message="Here is your report:")
If the user uploads an image (e.g., of a machine):
log.py script as usual.body_part from the exercise name (e.g., "Bench Press" -> "Chest") without asking, unless ambiguous.render.py.figsize width ~6).0 between drop set segments.Data is stored in skills/workout-tracker/workout.db (SQLite). The schema includes date, exercise, weight, unit, sets, reps, rest_time, notes.
This skill uses uv for Python package management.
Ensure uv is installed: brew install uv
Sync dependencies:
cd skills/workout-tracker/scripts
uv sync
The SQLite database (workout.db) uses the following schema:
id: Integer (PK)date: ISO8601 Stringexercise: Stringbody_part: String (Target muscle group)weight: Floatunit: String (default "kg")sets: Integerreps: Integerrest_time: Integer (Seconds, optional)notes: String (Optional)Weekly Installs
50
Repository
First Seen
Jan 30, 2026
Security Audits
Gen Agent Trust HubPassSocketPassSnykPass
Installed on
gemini-cli47
codex47
opencode47
github-copilot46
cursor46
cline45
Skills CLI 使用指南:AI Agent 技能包管理器安装与管理教程
52,700 周安装