360 Secure SkillHub
360 安全技能中心

running tracker

核心功能：解析用户自然语言表述的跑步数据，存储至指定文件；计算配速、速度、卡路里，生成与过往数据对比的进度反馈；支持查询跑步历史，计算均值、最佳成绩等，可合理外推未跑距离的成绩。典型场景：用户报告新跑步记录、询问配速/PB/训练建议等。输入为自然语言，输出是统计数据、分析结论或查询结果。

fdocr
fdocr
热度 28.33
最近更新 2026-05-11 17:48:12
分类：数据依赖：读写文件 / 代码执行 / 内部调用大模型应用场景：数据处理 / 健康 / 报告撰写

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add running-tracker
Metadata
name
running_tracker
description
Track, log, and analyze running workout times. Use when the user reports a new run (e.g. "1mi 8:20", "3k 15:33", "just finished a run"), asks about running history, pace, progress, personal bests, training advice based on their logged runs, or anything related to their running data.

Running Tracker

Log runs and analyze running performance using the data file at {baseDir}/runs.md.

Logging a New Run

Input parsing

Runs arrive in casual formats. Extract three fields:

  • Distance: 1mi, 3k, 5k, 10k, 1.5mi, etc.
  • Date: DD/MM/YY or natural language ("today", "yesterday"). If omitted, use today's date.
  • Time: M:SS or MM:SS (duration to complete the distance).

Storage

Append a new row to the markdown table in {baseDir}/runs.md. Store dates as YYYY-MM-DD. Keep the table sorted by date ascending (newest at the bottom).

Response after logging

  1. Run stats — compute and display:

    • Pace (min/km)
    • Speed (km/h)
    • Estimated calories burned (use 62 cal/km, no elevation)

  2. Progress note — 2-3 sentences comparing this run to recent history. Examples: pace trend, personal best alert, slowest/fastest in N days, streak observations. Be honest — if they slowed down, say so encouragingly.

Answering History Questions

Read {baseDir}/runs.md and compute whatever the user asks: averages, bests, trends, comparisons across distances, weekly/monthly summaries, training advice, etc.

Unit conversions

  • 1 mile = 1.60934 km
  • Pace = total minutes / distance in km
  • Speed = distance in km / (time in hours)

When the user's question involves a distance they haven't run (e.g. 10k projections), extrapolate cautiously and note the assumption.