marswangyang/workout-tracker

Overview

This skill logs and tracks strength and conditioning workouts using a local SQLite database and lightweight Python scripts. It captures sets, reps, weight, rest time and notes, and can auto-fill rest intervals from past sessions. It also analyzes images to identify equipment and offer quick logging suggestions.

How this skill works

The skill parses natural-language workout inputs and runs small scripts to insert structured records into a local SQLite database (date, exercise, body_part, weight, unit, sets, reps, rest_time, notes). If rest_time is omitted it infers a sensible default from the user’s recent history. When given an image, the vision component recognizes equipment, suggests the exercise, and prompts the user to confirm weight/sets before logging.

When to use it

• Quickly log a completed set or full workout from a short message.
• Check progress for a specific exercise or view recent sessions with a visual report.
• Upload a photo of a machine or dumbbells to identify equipment and log a set.
• Let the skill infer rest times based on your prior sessions when you don’t specify them.
• Track drop sets as separate rows for accurate analytics.

Best practices

• Provide exercise name, weight and reps in one phrase (e.g., “5x8 Bench Press 135lb”) to minimize follow-ups.
• Omit rest_time if you want the system to auto-fill from history; only specify it when deviating from your norm.
• Prefer pounds (lb) for units—this profile defaults to lb but accepts kg if specified.
• Log drop sets as separate entries; the system will set rest_time to 0 and mark notes as “Drop set part X/Y.”
• Upload clear photos of equipment (single item, well-lit) for reliable visual identification.

Example use cases

• User: “I did 4 sets of Squat at 225lb for 5 reps” → skill parses and logs the workout with inferred body_part = Legs.
• User uploads a picture of a leg press machine → skill identifies it, asks for weight, then logs the set on confirmation.
• User: “Show my Bench Press progress” → skill generates a mobile-optimized portrait report image summarizing historical reps/weights and sends it.
• User performs drop sets and reports segments separately → each segment is stored as its own row with rest_time=0 and an explanatory note.