mcpbeat Sign in

Healthkit Agent Skill

Read, write, and query Apple Health data using HealthKit. Covers HKHealthStore authorization, sample queries, statistics queries, statistics collection queries for charts, saving HKQuantitySample data, background delivery, workout sessions with HKWorkoutSession and HKLiveWorkoutBuilder, HKUnit, and HKQuantityTypeIdentifier values. Use when integrating with Apple Health, displaying health metrics, recording workouts, or enabling background health data delivery.

11k tokens
context cost
the whole folder, loaded on every use
3
files
instructions only
0
copies elsewhere
how many repositories repackaged it
960
stars on the repo
on the repository, not the skill itself

Install

one command, takes just this skill from the repository
npx skills add https://github.com/dpearson2699/swift-ios-skills --skill healthkit

The instruction itself

22 sections, as written by the author

HealthKit

Read and write health and fitness data from the Apple Health store. Covers authorization, queries, writing samples, background delivery, and workout sessions. Targets Swift 6.3 / iOS 26+.

Contents

  • Setup and Availability
  • Authorization
  • Reading Data: Sample Queries
  • Reading Data: Statistics Queries
  • Reading Data: Statistics Collection Queries
  • Writing Data
  • Background Delivery
  • Workout Sessions
  • Common Data Types
  • HKUnit Reference
  • Common Mistakes
  • Review Checklist
  • References

Setup and Availability

Project Configuration

  • Enable the HealthKit capability in Xcode (adds the entitlement)
  • Add NSHealthShareUsageDescription (read) and NSHealthUpdateUsageDescription (write) to Info.plist
  • For background delivery, enable the "Background Delivery" sub-capability

Availability Check

Always check availability before calling other HealthKit APIs. Health data is

available on iOS, watchOS, visionOS, iPadOS 17+, and iOS apps running on

Vision Pro. It is unavailable on iPadOS 16 or earlier and may be restricted by

managed device policy.

import HealthKit

guard HKHealthStore.isHealthDataAvailable() else {
    // Health data is unavailable or restricted on this device.
    return
}

let healthStore = HKHealthStore()

Create a single HKHealthStore instance and reuse it throughout your app. It

is thread-safe. If HealthKit is optional, review Xcode's generated

UIRequiredDeviceCapabilities healthkit entry so unsupported devices are not

excluded unintentionally.

Authorization

Request only the types your app genuinely needs. App Review rejects apps that over-request.

func requestAuthorization() async throws {
    let typesToShare: Set<HKSampleType> = [
        HKQuantityType(.stepCount),
        HKQuantityType(.activeEnergyBurned)
    ]

    let typesToRead: Set<HKObjectType> = [
        HKQuantityType(.stepCount),
        HKQuantityType(.heartRate),
        HKQuantityType(.activeEnergyBurned),
        HKCharacteristicType(.dateOfBirth)
    ]

    try await healthStore.requestAuthorization(
        toShare: typesToShare,
        read: typesToRead
    )
}

Checking Authorization Status

authorizationStatus(for:) reports write/share authorization. HealthKit does

not reveal whether read permission was granted or denied. If the user denies

read access, queries return only samples your app successfully saved, which may

look like empty or partial data.

let status = healthStore.authorizationStatus(
    for: HKQuantityType(.stepCount)
)

switch status {
case .notDetermined:
    // Haven't requested yet -- safe to call requestAuthorization
    break
case .sharingAuthorized:
    // User granted write access
    break
case .sharingDenied:
    // User denied write access (read denial is indistinguishable from "no data")
    break
@unknown default:
    break
}

Reading Data: Sample Queries

Use HKSampleQueryDescriptor (async/await) for one-shot reads. Prefer descriptors over the older callback-based HKSampleQuery.

func fetchRecentHeartRates() async throws -> [HKQuantitySample] {
    let heartRateType = HKQuantityType(.heartRate)

    let descriptor = HKSampleQueryDescriptor(
        predicates: [.quantitySample(type: heartRateType)],
        sortDescriptors: [SortDescriptor(\.endDate, order: .reverse)],
        limit: 20
    )

    let results = try await descriptor.result(for: healthStore)
    return results
}

// Extracting values from samples:
for sample in results {
    let bpm = sample.quantity.doubleValue(
        for: HKUnit.count().unitDivided(by: .minute())
    )
    print("\(bpm) bpm at \(sample.endDate)")
}

Reading Data: Statistics Queries

Use HKStatisticsQueryDescriptor for aggregated single-value stats (sum, average, min, max).

func fetchTodayStepCount() async throws -> Double? {
    let calendar = Calendar.current
    let startOfDay = calendar.startOfDay(for: Date())
    let endOfDay = calendar.date(byAdding: .day, value: 1, to: startOfDay)!

    let predicate = HKQuery.predicateForSamples(
        withStart: startOfDay, end: endOfDay
    )
    let stepType = HKQuantityType(.stepCount)
    let samplePredicate = HKSamplePredicate.quantitySample(
        type: stepType, predicate: predicate
    )

    let query = HKStatisticsQueryDescriptor(
        predicate: samplePredicate,
        options: .cumulativeSum
    )

    let result = try await query.result(for: healthStore)
    return result?.sumQuantity()?.doubleValue(for: .count())
}

Options by data type:

  • Cumulative types (steps, calories): .cumulativeSum
  • Discrete types (heart rate, weight): .discreteAverage, .discreteMin, .discreteMax

Reading Data: Statistics Collection Queries

Use HKStatisticsCollectionQueryDescriptor for time-series data grouped into intervals -- ideal for charts.

func fetchDailySteps(forLast days: Int) async throws -> [(date: Date, steps: Double)] {
    let calendar = Calendar.current
    let endDate = calendar.startOfDay(
        for: calendar.date(byAdding: .day, value: 1, to: Date())!
    )
    let startDate = calendar.date(byAdding: .day, value: -days, to: endDate)!

    let predicate = HKQuery.predicateForSamples(
        withStart: startDate, end: endDate
    )
    let stepType = HKQuantityType(.stepCount)
    let samplePredicate = HKSamplePredicate.quantitySample(
        type: stepType, predicate: predicate
    )

    let query = HKStatisticsCollectionQueryDescriptor(
        predicate: samplePredicate,
        options: .cumulativeSum,
        anchorDate: endDate,
        intervalComponents: DateComponents(day: 1)
    )

    let collection = try await query.result(for: healthStore)
    var dailySteps: [(date: Date, steps: Double)] = []

    collection.statisticsCollection.enumerateStatistics(
        from: startDate, to: endDate
    ) { statistics, _ in
        let steps = statistics.sumQuantity()?
            .doubleValue(for: .count()) ?? 0
        dailySteps.append((date: statistics.startDate, steps: steps))
    }

    return dailySteps
}

Long-Running Collection Query

Use results(for:) (plural) to get an AsyncSequence that emits updates as new data arrives:

let updateStream = query.results(for: healthStore)

Task {
    for try await result in updateStream {
        // result.statisticsCollection contains updated data
    }
}

Writing Data

Create HKQuantitySample objects and save them to the store.

func saveSteps(count: Double, start: Date, end: Date) async throws {
    let stepType = HKQuantityType(.stepCount)
    let quantity = HKQuantity(unit: .count(), doubleValue: count)

    let sample = HKQuantitySample(
        type: stepType,
        quantity: quantity,
        start: start,
        end: end
    )

    try await healthStore.save(sample)
}

Treat try await healthStore.save(sample) returning as the save success gate;

only then report success or advance app state. On failure, surface the error and

correct the known authorization, type, unit, duration, or input problem before

constructing another sample. A bounded query or inspection in the Health app is

useful as an integration-test check when persistence evidence is required, but

is not a mandatory production read after every save.

Your app can only delete samples it created. Samples from other apps or Apple Watch are read-only.

Background Delivery

Register for background updates so your app is launched when new data arrives. Requires the background delivery entitlement.

func enableStepCountBackgroundDelivery() async throws {
    let stepType = HKQuantityType(.stepCount)

    try await healthStore.enableBackgroundDelivery(
        for: stepType,
        frequency: .hourly
    )
}

Pair with an HKObserverQuery to handle notifications. Always call the completion handler:

let observerQuery = HKObserverQuery(
    sampleType: HKQuantityType(.stepCount),
    predicate: nil
) { query, completionHandler, error in
    defer { completionHandler() }  // Must call to signal done
    guard error == nil else { return }
    // Fetch new data, update UI, etc.
}
healthStore.execute(observerQuery)

Frequencies: .immediate, .hourly, .daily, .weekly

Set up observer queries as soon as the app launches, then call

enableBackgroundDelivery once for the same sample type. The system persists

the registration, wakes the app at most once per requested frequency, and

enforces tighter caps for some types such as hourly step-count delivery on iOS.

Background delivery is not supported on Simulator; test it on device.

Workout Sessions

Use HKWorkoutSession and HKLiveWorkoutBuilder to track live workouts.

HKWorkoutSession is available on iOS/iPadOS 17+, visionOS 1+, and watchOS 2+.

HKLiveWorkoutBuilder is available on iOS/iPadOS 26+ and watchOS 5+, so gate

live-builder code if supporting older iOS/iPadOS releases.

On iPhone and iPad, live heart-rate collection requires a paired external heart

rate sensor. Apple Watch sessions can collect high-frequency heart-rate data.

For locked iPhone workouts, plan for the system's workout-data access flow

before showing health metrics on the Lock Screen.

func startWorkout() async throws {
    let configuration = HKWorkoutConfiguration()
    configuration.activityType = .running
    configuration.locationType = .outdoor

    let session = try HKWorkoutSession(
        healthStore: healthStore,
        configuration: configuration
    )
    session.delegate = self

    let builder = session.associatedWorkoutBuilder()
    builder.dataSource = HKLiveWorkoutDataSource(
        healthStore: healthStore,
        workoutConfiguration: configuration
    )

    session.startActivity(with: Date())
    try await builder.beginCollection(at: Date())
}

// Request teardown; finalize from the delegate's .stopped transition.
session.stopActivity(with: Date())

Do not call endCollection and finishWorkout immediately after requesting the

stop. Wait for the session delegate's .stopped transition, then await

builder.endCollection(at:) followed by builder.finishWorkout(). Report the

workout as saved and clear session state only after both operations return.

Handle each thrown error without blindly repeating teardown. A successful

finishWorkout() can return no workout object while the device is locked, so a

nil result alone is not failure.

For full workout lifecycle management including pause/resume, delegate handling, and multi-device mirroring, see references/healthkit-patterns.md.

Common Data Types

HKQuantityTypeIdentifier

| Identifier | Category | Unit |

|---|---|---|

| .stepCount | Fitness | .count() |

| .distanceWalkingRunning | Fitness | .meter() |

| .activeEnergyBurned | Fitness | .kilocalorie() |

| .basalEnergyBurned | Fitness | .kilocalorie() |

| .heartRate | Vitals | .count()/.minute() |

| .restingHeartRate | Vitals | .count()/.minute() |

| .oxygenSaturation | Vitals | .percent() |

| .bodyMass | Body | .gramUnit(with: .kilo) |

| .bodyMassIndex | Body | .count() |

| .height | Body | .meter() |

| .bodyFatPercentage | Body | .percent() |

| .bloodGlucose | Lab | .gramUnit(with: .milli).unitDivided(by: .literUnit(with: .deci)) |

HKCategoryTypeIdentifier

Common category types: .sleepAnalysis, .mindfulSession, .appleStandHour

HKCharacteristicType

Read-only user characteristics include .dateOfBirth, .biologicalSex,

.bloodType, .fitzpatrickSkinType, .wheelchairUse, and .activityMoveMode.

HKUnit Reference

// Basic units
HKUnit.count()                              // Steps, counts
HKUnit.meter()                              // Distance
HKUnit.mile()                               // Distance (imperial)
HKUnit.kilocalorie()                        // Energy
HKUnit.joule(with: .kilo)                   // Energy (SI)
HKUnit.gramUnit(with: .kilo)                // Mass (kg)
HKUnit.pound()                              // Mass (imperial)
HKUnit.percent()                            // Percentage

// Compound units
HKUnit.count().unitDivided(by: .minute())   // Heart rate (bpm)
HKUnit.meter().unitDivided(by: .second())   // Speed (m/s)

// Prefixed units
HKUnit.gramUnit(with: .milli)               // Milligrams
HKUnit.literUnit(with: .deci)               // Deciliters

Common Mistakes

  • Over-requesting data types. Request only the read/write types the feature

actually uses; broad HealthKit permission sheets are an App Review risk.

  • Treating read authorization like write authorization. You can check

.sharingAuthorized before saving, but read denial is privacy-protected and

looks like app-owned-only, empty, or partial results.

  • Skipping isHealthDataAvailable(). Check before HealthKit access and

handle unavailable or restricted stores without crashing.

  • Using callback queries for new async code. Prefer async descriptors for

one-shot reads and statistics, and keep broad queries off the main actor.

  • Forgetting observer completion handlers. Always call the handler; missed

completions can delay or stop future background deliveries.

  • Assuming .immediate means immediate. Background delivery is capped by

the system and must be tested on device.

  • Using cumulative stats for discrete values. Match statistics options to

the data type: cumulative sums for steps/energy, discrete average/min/max for

heart rate, weight, and similar samples.

Review Checklist

  • [ ] HKHealthStore.isHealthDataAvailable() checked before any HealthKit access
  • [ ] Only necessary data types requested in authorization
  • [ ] Info.plist includes NSHealthShareUsageDescription and/or NSHealthUpdateUsageDescription
  • [ ] HealthKit capability enabled in Xcode project
  • [ ] Write authorization checked before saving; read denial handled as partial

or empty query results

  • [ ] Single HKHealthStore instance reused (not created per query)
  • [ ] Async query descriptors used instead of callback-based queries
  • [ ] Heavy queries not blocking main thread
  • [ ] Statistics options match data type (cumulative vs. discrete)
  • [ ] Background delivery paired with app-launch HKObserverQuery setup and

completionHandler called

  • [ ] Background delivery entitlement enabled if using enableBackgroundDelivery
  • [ ] Background delivery tested on device and frequency caps considered
  • [ ] Workout stop waits for the delegate's .stopped transition before

endCollection and finishWorkout; state clears only after successful

finalization

  • [ ] Workout API availability and live heart-rate sensor requirements handled
  • [ ] Delete operations target only objects the app previously saved

References

Other skills for the same job

different authors, same section of the catalogue
XLSX
by anthropics
vendor ×15

Comprehensive spreadsheet creation, editing, and analysis with support for formulas, formatting, data analysis, and visualization. When Claude needs to work with spreadsheets (.xlsx, .xlsm, .csv, .tsv, etc) for: (1) Creating new spreadsheets with formulas and formatting, (2) Reading or analyzing data, (3) Modify existing spreadsheets while preserving formulas, (4) Data analysis and visualization in spreadsheets, or (5) Recalculating formulas

5k tokens scripts
XLSX
by w95
×7

Use this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like \"the xlsx in my downloads\") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved.

3k tokens
Raffle Winner Picker
by frostant
×5

Picks random winners from lists, spreadsheets, or Google Sheets for giveaways, raffles, and contests. Ensures fair, unbiased selection with transparency.

949 tokens
Fda Database
by christophacham
×4

Query openFDA API for drugs, devices, adverse events, recalls, regulatory submissions (510k, PMA), substance identification (UNII), for FDA regulatory data analysis and safety research.

32k tokens scripts
Matlab
by christophacham
×4

MATLAB and GNU Octave numerical computing for matrix operations, data analysis, visualization, and scientific computing. Use when writing MATLAB/Octave scripts for linear algebra, signal processing, image processing, differential equations, optimization, statistics, or creating scientific visualizations. Also use when the user needs help with MATLAB syntax, functions, or wants to convert between MATLAB and Python code. Scripts can be executed with MATLAB or the open-source GNU Octave interpreter.

25k tokens
Umap Learn
by ComeOnOliver
×4

UMAP dimensionality reduction. Fast nonlinear manifold learning for 2D/3D visualization, clustering preprocessing (HDBSCAN), supervised/parametric UMAP, for high-dimensional data.

14k tokens
D3 Viz
by chrisvoncsefalvay
×3

Creating interactive data visualisations using d3.js. This skill should be used when creating custom charts, graphs, network diagrams, geographic visualisations, or any complex SVG-based data visualisation that requires fine-grained control over visual elements, transitions, or interactions. Use this for bespoke visualisations beyond standard charting libraries, whether in React, Vue, Svelte, vanilla JavaScript, or any other environment.

20k tokens
Alphafold Database
by christophacham
×3

Access AlphaFold 200M+ AI-predicted protein structures. Retrieve structures by UniProt ID, download PDB/mmCIF files, analyze confidence metrics (pLDDT, PAE), for drug discovery and structural biology.

7k tokens

How to use it

Copy the folder

Take dpearson2699/healthkit from the repository into ~/.claude/skills for personal use, or into .claude/skills inside a project.

Check the name does not clash

The agent identifies a skill by the name field in its header. Two skills with the same name cannot sit side by side — one of them will be ignored.