What problem does it solve? Building integrations that deliver structured workouts to users' fitness devices requires handling nine different providers (Garmin, COROS, Wahoo, Suunto, TrainingPeaks, Huawei, Zepp, Hevy, Apple), each with different feature support, operations, and quirks. This Skill gives an AI agent the endpoint reference, data model, and per-provider behavior needed to write correct Terra API Planned Workouts integration code. ## Core Features & Use Cases - Two-phase workflow guidance: Create a reusable workout template with POST /workouts, then schedule it per athlete with POST /workouts/{id}/plan?user_id=X, which applies athlete parameters (FTP, max HR, threshold values) and pushes to the device. - Data model and validation rules: Covers blocks, steps, completion conditions, and intensity targets (absolute, percentage, zone), including hard 400 rejections for out-of-range values and single-value restrictions on percentage targets. - Coercion warning handling: Explains how the API adapts payloads when a provider cannot represent a feature, with a full warning catalog and handling patterns (log, surface, ignore). - Provider compatibility matrices: Operations support, sport/target/completion support, and per-provider gotchas such as Zepp's 7-day sync window, Huawei's create-only running limitation, and COROS delete-and-recreate updates. - Use Case: A developer building a coaching app asks the agent to schedule a threshold-interval cycling workout personalized to each athlete's FTP; the agent produces the correct template and plan requests and anticipates coercion warnings on the target device. ## Quick Start Ask the agent to create a Terra API workout template for threshold intervals and schedule it to a user with a given FTP and planned date.
