\r \r

clawhealth-garmin (OpenClaw Skill)\r

\r Sync your Garmin Connect health data into a local SQLite database\r and expose it as structured JSON for OpenClaw agents.\r \r Your agent can then query things like:\r \r

• "How did I sleep yesterday?"\r
• "What is my HRV trend this week?"\r
• "Am I overtraining?"\r \r This is a thin wrapper skill around the published clawhealth\r Python package / CLI; it no longer fetches source code itself.\r \r

What It Does\r

\r

• Login with username/password (MFA supported)\r
• Sync daily health summaries into SQLite (stage 1)\r
• Fetch HRV + training metrics via separate commands (stage 2)\r
• Fetch sleep stages + sleep score (stage 2)\r
• Fetch body composition (stage 2)\r
• Fetch activity lists and full activity details (stage 2)\r
• Fetch menstrual day view and calendar range if supported by garminconnect (experimental)\r
• Provide --json outputs for agent workflows\r
• Persist raw JSON payloads for later analysis\r \r

Prerequisites\r

\r

• Python 3.10+\r
• Network access to Garmin Connect\r
• Garmin account (may require MFA)\r \r If you run OpenClaw in Docker, you may prefer a prepatched image that already\r includes the required Python dependencies:\r \r
• ernestyu/openclaw-patched\r \r

Setup\r

\r

1. Create {baseDir}/.env (see {baseDir}/ENV_EXAMPLE.md).\r \r Recommended: use CLAWHEALTH_GARMIN_PASSWORD_FILE (password file) rather than\r CLAWHEALTH_GARMIN_PASSWORD (plaintext env var).\r \r Note: relative paths in env vars (like ./garmin_pass.txt) are resolved relative\r to the skill directory by run_clawhealth.py.\r \r
2. Install the clawhealth package into a local .venv (if needed):\r \r

python {baseDir}/bootstrap_deps.py\r
```\r
\r
3) Run the skill entrypoint via OpenClaw, which will invoke `run_clawhealth.py`.\r
\r
See `README.md` / `README_zh.md` in this directory and the root repo for\r
more details.\r