Apple Healthkit logo

Apple Healthkit

Community
win4r
apple_healthkit

The user's synced Apple Health (HealthKit) data: daily metrics (steps, distance, calories, heart rate, HRV, VO2max), sleep sessions (stages, quality, efficiency), and workouts.

Overview

Publisherwin4r
RepositoryMuseAI-Skills
Skill nameapple_healthkit
Stars
382
Forks
116
Bundled files
1
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 1 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by win4r on GitHub. Read the source before you install it.

Installation

Install the Apple Healthkit AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/win4r/MuseAI-Skills/tree/main/opt/hatch/skills/apple-healthkit
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/win4r/MuseAI-Skills.git /tmp/MuseAI-Skills
mkdir -p .claude/skills
cp -r /tmp/MuseAI-Skills/opt/hatch/skills/apple-healthkit .claude/skills/apple_healthkit
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Apple Healthkit in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Apple Healthkit on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Apple Healthkit is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Apple Health

Read the user's synced Apple HealthKit data with the health-cli binary. Every command below takes --provider healthkit.

This skill serves iOS devices only. If the user's paired device is not an iPhone, it has no data for them — check the platform field returned by device.list when you are unsure.

Data Source

Despite syned from apple healthkit, do not assume the data is originated from ios devices only. It could come from other sources/devices which share data with Apple Healthkit.

When to Use

When the user asks about their synced HealthKit data, sync status, or sources:

  • All-day metrics/vitals: steps, distance, calories, heart rate, HRV, VO2max.
  • Sleep sessions: stages, quality, efficiency, awakenings.
  • Workouts: type, duration, calories, distance, HR.

Tooling

Binary: health-cli. --provider healthkit is required on every command below — it is not optional and has no default; omitting it is a usage error. Commands group by data shape under query:

  • query metrics --provider healthkit — all-day metric rollups
  • query samples --provider healthkit — raw, unbucketed data points (intraday HR/steps, GPS, sleep stages)
  • query sessions --provider healthkit — discrete sessions: sleep, workout
  • status --provider healthkit — data-sync status
  • auth connect --provider healthkit, auth disconnect --provider healthkit
  • delete — erase stored records (destructive; see below)

Multi-device (rare): when a window spans >1 device the envelope adds "multi_node": true — query metrics returns per-device node_groups; query samples/sessions add a node_id to each record (samples CSV gains a leading node_id column). Single-device output is unchanged; never sum or double-count across devices.

query metrics — all-day metric rollups

bash
health-cli query metrics --provider healthkit \
  --start-date <YYYY-MM-DD> [--end-date <YYYY-MM-DD>] \
  [--interval hourly|daily|weekly] [--fields a,b,c] [--timeout-secs N]

daily-metrics is the only domain so there is no --category. One row per --interval bucket (default daily). --start-date required; --end-date defaults to today. Output is an envelope { "coverage": {...}, "records": [...] } (see Coverage below).

--list-fields reads the metric field names from the synced data → { ok, provider, category, observed_count, fields: [{ name, observed, count? }] }. observed: true (with a row count) means the name is present in this user's data — those are exactly the names --fields matches. observed: false means the build knows the field but nothing has synced yet. If the field set can't be read the response carries degraded: true and every observed is null.

query sessions — discrete sessions

bash
health-cli query sessions --provider healthkit --category sleep|workout \
  --start-date <YYYY-MM-DD> [--end-date <YYYY-MM-DD>] [--fields a,b,c]
  • sleep — one row per sleep session.
  • workout — one row per workout/activity.

Sessions are never bucketed (no --interval). Discovery: --list-categories → { ok, provider, categories }; --list-fields (needs --category) → { ok, provider, category, fields } (e.g. workout has is_indoor).

Output: normalized snake_case records; the same concept uses the same field name (e.g. workout average_speed_mps, elevation_gain_meters, hr_average_bpm). Session records carry a provider-prefixed id (healthkit_…) + start_datetime/end_datetime/timezone; metric buckets carry date/hour/week_start + record_count.

Coverage: queries return an envelope { "coverage": {...}, "records": [...] }. coverage.complete is false when days in the window aren't synced from the device; coverage.warning then holds the exact backfill_data_source command to run. Surface that warning to the user / act on it — results are partial until the backfill completes.

query samples — raw, unbucketed data points

bash
health-cli query samples --provider healthkit --start-date <YYYY-MM-DD> \
  [--end-date <YYYY-MM-DD>] [--start-time <HH:MM[:SS]>] [--end-time <HH:MM[:SS]>] \
  [--fields <type1,...>] [--limit <n>] [--format stdout|csv] [--output <path>]
health-cli query samples --provider healthkit --list-fields   # discover sample types

Individual samples — finer than the query metrics rollups (intraday HR/steps, GPS, sleep-stage timelines; query sessions gives stage totals only). --fields selects sample types (not output columns) — run --list-fields to see the types this user has synced. GPS is location (raw lat/lon — don't infer place names from it). For dense reads use --output <path>: it writes CSV into the agent filesystem for a later script step and prints only a summary, keeping thousands of rows out of context.

status — data-sync status

bash
health-cli status --provider healthkit [--timeout-secs N] \
  [--start-date <YYYY-MM-DD>] [--end-date <YYYY-MM-DD>] [--check-missing-entries]

Returns { categories: [{ name, record_count, earliest_datetime, latest_datetime }] }. With --check-missing-entries (requires --start-date): a deep per-30-min-interval gap report (unsynced_dates, missing_intervals, …).

auth connect

bash
health-cli auth connect --provider healthkit [--timeout-secs N]

Run this and follow the instruction to connect to apple healthkit data.

auth disconnect

bash
health-cli auth disconnect --provider healthkit [--timeout-secs N]

Apple Health cannot be disconnected through this CLI or an in-chat widget. Tell the user to open Settings → Connectors → Apple Health and tap Disconnect.

delete — erase stored Apple Health (HealthKit) records

bash
health-cli delete --provider healthkit --all                # every Apple Health record
health-cli delete --provider healthkit --device NODE_ID     # one paired device
health-cli delete --provider healthkit --start-date 2026-03-01 --end-date 2026-03-31

Use for any request to delete, remove, erase, wipe, or clear HealthKit data. Never improvise it by moving files, disconnecting the connector, or writing to the database — those neither remove records nor leave an audit trail. A delete never spans providers; to clear both, run it once per provider.

--all takes no other filter; --device/--start-date/--end-date are used instead of it. All record kinds go at once — no per-category or per-record-ID delete. Dates are YYYY-MM-DD in JARVIS_USER_TIMEZONE or epoch seconds; a record is in range when it starts before the end and ends after the start, so a session straddling a bound is included. --start-date alone deletes from that day onward, --end-date alone up to and including it.

Destructive and irreversible, behind a fresh one-time approval showing the normalized filters. Run it once per request — settle provider, device and dates in conversation first. Denial exits code 3 with nothing deleted; that is final, so report it rather than retrying other filters or the other provider.

Zero is a success, not a miss — nothing matched, so say so and stop rather than widening dates or re-running with --all. Report deleted.records, not total_rows, which also counts child value rows. Apple Health is device-synced, so records still on the device can return on a later sync.

Auth

Device-synced; no login. Data availability depends on the on-device Health permission + sync. Use status to see synced ranges and the backfill_data_source device action (surfaced in query coverage.warning) to pull a range.

Operating Rules

  1. Pick the command by data shape: query metrics (all-day rollups), query samples (raw points — only when individual samples matter; prefer metrics for trends), or query sessions --category sleep|workout (discrete events).
  2. Always pass --provider healthkit and --start-date; --end-date defaults to today. Use current-date context for relative asks ("today", "last week"); use sample time bounds for narrow intraday windows.
  3. Coverage: if a query metrics/sessions returns coverage.complete: false, tell the user the data is partial and relay/act on coverage.warning (run the backfill_data_source action, then re-query).
  4. query metrics defaults to --interval daily; sleep spans midnight (include both evening start and morning end dates).
  5. On empty results, the range likely has no data; widen it, or check status / coverage and backfill.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Apple Healthkit AI skill do?

The user's synced Apple Health (HealthKit) data: daily metrics (steps, distance, calories, heart rate, HRV, VO2max), sleep sessions (stages, quality, efficiency), and workouts.

Why use Apple Healthkit on TypingMind?

Because you install it once and use it with any model. Apple Healthkit is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Apple Healthkit in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/win4r/MuseAI-Skills/tree/main/opt/hatch/skills/apple-healthkit. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Apple Healthkit?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Apple Healthkit?

As many as you like. As long as a model supports skills, you can use Apple Healthkit with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Apple Healthkit AI skill free?

It is published on GitHub by win4r. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇