add-exercise_skill

This skill helps you add a new exercise to vibereps by generating a JSON config and determining the appropriate detection type.
  • Python

1

GitHub Stars

1

Bundled Files

2 months ago

Catalog Refreshed

4 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill flow-club/vibereps --skill add-exercise

  • SKILL.md1.8 KB

Overview

This skill adds a new exercise type to vibereps by generating the JSON config and the corresponding pose-detection logic. It streamlines creating detection rules, thresholds, and user instructions so the exercise is trackable by the tracker. It’s designed for quick, repeatable addition of common motion patterns.

How this skill works

The skill inspects chosen MediaPipe landmark IDs and selects a detection type (angle, height_baseline, height_relative, tilt, distance, width_ratio, quadrant_tracking) to match the motion. It builds a JSON config following the template structure and injects landmark arrays and threshold values. Finally, it validates the config by running the tracker so detection behavior can be tested and refined.

When to use it

  • You want to add a new exercise not yet in the app.
  • You need pose-based repetition detection for a custom move.
  • You are defining thresholds or alternate joint sets for left/right symmetry.
  • You want a fast prototype for testing detection logic with the tracker.
  • You need clear on-screen instructions and rep presets for an exercise.

Best practices

  • Pick the detection type that matches the primary motion (angle for joint flexion, height for vertical moves, distance for parts approaching).
  • Provide both left and right landmark sets when motion can be asymmetric.
  • Set conservative thresholds, test live, then tighten values to reduce false positives.
  • Write concise user instructions tied to threshold tokens (e.g., {down}, {up}).
  • Run the tracker after each edit and iterate using real movement to validate detection.

Example use cases

  • Add a squat using 'angle' detection with hip-knee-ankle joint arrays and down/up angle thresholds.
  • Create calf raises using 'height_baseline' to measure ankle vertical movement from a baseline.
  • Implement jumping jacks via 'height_relative' to detect hands moving above a reference point.
  • Build torso twists using 'width_ratio' to monitor shoulder/hip width changes.
  • Define arm circles with 'quadrant_tracking' to follow circular wrist motion.

FAQ

Match the type to the dominant motion: use angle for joint flexion, height types for vertical movement, distance for parts approaching, width_ratio for lateral spread, tilt for torso lean, and quadrant_tracking for circular limb paths.

How can I test and refine thresholds?

Create the JSON config, run the tracker with live movement, observe detection hits/misses, then iteratively adjust threshold values and retest.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational