← Back to Skills Marketplace
kinsonhu123-mkll

AI Running Coach

by kinsonhu123-mkll · GitHub ↗ · v1.0.1 · MIT-0
cross-platform ✓ Security Clean
217
Downloads
1
Stars
0
Active Installs
2
Versions
Install in OpenClaw
/install airunningcoach
Description
AI 跑步教练 - 基于 VDOT 理论和周期化训练的专业马拉松训练指导。使用场景:制定马拉松/半马训练计划、分析跑步数据(TCX/GPX)、计算配速/心率区间、比赛策略制定、训练问题咨询、伤病预防建议。
README (SKILL.md)

AI 跑步教练技能

专业的马拉松训练指导,基于 Jack Daniels VDOT 理论和极化训练原则。


核心功能

功能 说明 触发示例
训练计划制定 生成个性化周期化训练计划 "帮我制定破 3 训练计划"
配速/心率计算 基于 VDOT 计算各区间配速和心率 "破 3 需要什么配速"
跑步数据分析 解析 TCX/GPX 文件,分析训练表现 "分析我今天的跑步数据"
比赛策略制定 制定比赛配速和补给策略 "全马比赛怎么配速"
训练问题咨询 回答训练方法、伤病预防等问题 "间歇跑怎么练"
成绩预测 基于训练数据预测比赛成绩 "我能跑多少"

工作流程

1. 训练计划制定

触发: 用户提到"训练计划"、"马拉松"、"备赛"等

步骤:

  1. 收集信息(如未提供):

    • 当前 PB 或最近比赛成绩
    • 目标赛事和目标时间
    • 当前周跑量和训练频率
    • 年龄、体重(可选)
    • 伤病史(可选)
  2. 计算 VDOT

    python scripts/calculate_pace.py --race-time \x3C分钟数> --distance 42.195
    
  3. 生成计划

    python scripts/generate_plan.py --current-pb \x3CHH:MM> --target \x3CHH:MM> --weeks 24 --weekly-km \x3C跑量>
    
  4. 输出计划

    • 阶段划分(基础期、强化期、巅峰期、减量期)
    • 周训练安排(含配速、心率区间)
    • 关键训练课说明

参考: references/training-zones.mdreferences/workout-types.md


2. 配速/心率区间计算

触发: 用户询问配速、心率区间、VDOT 等

步骤:

  1. 确定输入

    • 如有比赛成绩:用成绩计算 VDOT
    • 如有 VDOT:直接使用
    • 如有最大心率/静息心率:计算心率区间
  2. 计算

    # 基于比赛成绩
    python scripts/calculate_pace.py --race-time 218 --distance 42.195 --max-hr 196 --resting-hr 55
    
    # 基于 VDOT
    python scripts/calculate_pace.py --vdot 44 --max-hr 196
    
  3. 输出

    • 各配速区间(E/M/T/I/R)
    • 各心率区间(Z1-Z5)
    • 对应训练用途说明

参考: references/training-zones.md


3. 跑步数据分析

触发: 用户上传 TCX/GPX 文件或询问训练数据分析

步骤:

  1. 解析文件

    python scripts/analyze_run.py \x3C文件路径> --output json
    
  2. 分析内容

    • 总距离、总时间、平均配速
    • 平均心率、最大心率
    • 平均步频
    • 心率区间分布
    • 配速稳定性
  3. 给出建议

    • 训练强度是否合适
    • 心率是否偏高/偏低
    • 改进建议

参考: references/training-zones.mdreferences/workout-types.md


4. 比赛策略制定

触发: 用户询问比赛配速、补给策略、赛前准备等

步骤:

  1. 收集信息

    • 目标赛事和目标时间
    • 当前训练水平(PB、最近训练配速)
    • 比赛地点和天气(可选)
  2. 制定策略

    • 配速策略(匀速/负分割)
    • 补给计划(胶、盐丸、水)
    • 赛前一周安排
    • 比赛日清单
  3. 输出

    • 各公里点目标时间
    • 补给时机
    • 心理策略

参考: references/race-strategies.md


5. 训练问题咨询

触发: 用户询问训练方法、伤病、装备等

处理:

  1. 判断问题类型

    • 训练方法 → 参考 references/workout-types.md
    • 心率/配速 → 参考 references/training-zones.md
    • 比赛策略 → 参考 references/race-strategies.md
  2. 给出专业回答

    • 基于跑步科学理论
    • 结合用户实际情况
    • 提供可执行建议

脚本说明

calculate_pace.py

功能: 配速和心率区间计算

参数:

  • --vdot:VDOT 值
  • --race-time:比赛时间(分钟)
  • --distance:比赛距离(公里,默认 42.195)
  • --max-hr:最大心率
  • --resting-hr:静息心率
  • --output:输出格式(json/text)

示例:

python scripts/calculate_pace.py --race-time 218 --distance 42.195 --max-hr 196 --output json

analyze_run.py

功能: TCX/GPX 数据分析

参数:

  • file:文件路径(必填)
  • --output:输出格式(json/text)

示例:

python scripts/analyze_run.py run.tcx --output json

generate_plan.py

功能: 马拉松训练计划生成

参数:

  • --current-pb:当前 PB(HH:MM 或分钟数)
  • --target:目标时间(HH:MM 或分钟数)
  • --weeks:训练周期(周,默认 24)
  • --weekly-km:当前周跑量(默认 300)
  • --output:输出格式(json/text)

示例:

python scripts/generate_plan.py --current-pb 3:38 --target 3:00 --weeks 24 --weekly-km 300

关键理论

VDOT 理论(Jack Daniels)

  • 基于比赛成绩量化跑步能力
  • 提供 5 个配速区间(E/M/T/I/R)
  • 科学指导训练强度

极化训练(80/20 法则)

  • 80% 低强度(Z1-Z2)
  • 20% 高强度(Z4-Z5)
  • 避免"垃圾跑量"

周期化训练

  • 基础期:有氧基础
  • 强化期:阈值 + 间歇
  • 巅峰期:马拉松配速 + 最长 LSD
  • 减量期:恢复 + 保持

注意事项

  1. 个体差异:公式计算是参考,需结合个人感受调整
  2. 循序渐进:周跑量增幅不超过 10%
  3. 恢复优先:感觉疲劳时优先休息
  4. 伤病预警:持续疼痛应停止训练并就医
  5. 数据局限:TCX/GPX 解析依赖文件格式,部分数据可能缺失

参考文档

  • references/training-zones.md - 训练区间定义
  • references/workout-types.md - 训练课类型详解
  • references/race-strategies.md - 比赛策略指南
Usage Guidance
This skill appears coherent and implements the described running-coach features locally. Before installing/using: 1) Ensure the agent environment has a trusted Python runtime (SKILL.md assumes 'python') since metadata didn't list required binaries. 2) Review the included scripts (they run locally and do not perform network calls) if you need to verify behavior. 3) Be aware you will provide personal/health data (race times, heart rates, possibly GPX/TCX files) — treat that as sensitive and avoid sending it to third parties. 4) Note some defaults (e.g., default weekly_km values) may be unrealistic; validate outputs and consult a qualified coach/medical professional for injury concerns. If you want extra assurance, run the scripts in a sandboxed environment or inspect/execute them locally before enabling autonomous invocation.
Capability Analysis
Type: OpenClaw Skill Name: airunningcoach Version: 1.0.1 The 'airunningcoach' skill bundle is a legitimate tool designed to provide marathon training guidance based on VDOT theory. The included Python scripts (analyze_run.py, calculate_pace.py, generate_plan.py) perform data parsing and mathematical calculations related to running metrics without any network activity, sensitive data access, or obfuscation. The SKILL.md instructions correctly guide the AI agent to use these scripts for their intended purposes, and no indicators of prompt injection or malicious intent were found.
Capability Assessment
Purpose & Capability
Name/description (VDOT-based running coach) matches the included scripts (plan generation, run analysis, pace/HR calculations) and reference docs. The functionality implemented is appropriate for the stated purpose.
Instruction Scope
SKILL.md instructs the agent to run the provided Python scripts and to accept user-uploaded TCX/GPX files and personal data (PB, target time, HR, optional age/weight/injury history). That behaviour is expected for this skill. Minor inconsistency: metadata did not declare Python as a required binary even though runtime examples call 'python'.
Install Mechanism
No install spec — the skill is delivered as scripts and documentation to run in-place. No downloads, external installers, or URLs are used. This is low-risk from an install mechanism perspective.
Credentials
The skill requests no environment variables, credentials, or config paths. It only processes user-supplied activity files and optional personal/physiological inputs, which are proportionate to the stated functionality.
Persistence & Privilege
always:false and default invocation settings — the skill does not request permanent/always-on privileges or modifications to other skills; no elevated platform privileges are requested.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install airunningcoach
  3. After installation, invoke the skill by name or use /airunningcoach
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.1
Initial release of AI Running Coach! - 提供基于 VDOT 理论和周期化训练的专业马拉松训练指导 - 支持个性化训练计划生成、配速和心率区间计算 - 能分析跑步数据(TCX/GPX)、并给出训练建议 - 帮助制定比赛配速与补给策略,以及训练/伤病等常见问题咨询 - 参考主流跑步科学理论,包括 VDOT、极化训练和周期化训练
v1.0.0
- Initial release of AI Running Coach based on Jack Daniels’ VDOT theory and periodized marathon training. - Provides personalized marathon and half-marathon training plans. - Supports pace/heart rate zone calculations, race strategy, and performance analysis from TCX/GPX files. - Offers training Q&A, injury prevention advice, and performance prediction. - Includes command-line tools for pace calculation, plan generation, and run data analysis.
Metadata
Slug airunningcoach
Version 1.0.1
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 2
Frequently Asked Questions

What is AI Running Coach?

AI 跑步教练 - 基于 VDOT 理论和周期化训练的专业马拉松训练指导。使用场景:制定马拉松/半马训练计划、分析跑步数据(TCX/GPX)、计算配速/心率区间、比赛策略制定、训练问题咨询、伤病预防建议。 It is an AI Agent Skill for Claude Code / OpenClaw, with 217 downloads so far.

How do I install AI Running Coach?

Run "/install airunningcoach" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is AI Running Coach free?

Yes, AI Running Coach is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does AI Running Coach support?

AI Running Coach is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created AI Running Coach?

It is built and maintained by kinsonhu123-mkll (@kinsonhu123-mkll); the current version is v1.0.1.

💬 Comments