agentleFS
mycelium-hqmycelium-hq/ai-brain-starter/skills/health-doctor/SKILL.md

health-doctor

Skill50 starsChanged 7 months agoMITView on GitHub

Use when the user runs /health doctor or /health status, asks whether the health auto-chain is working, when the last Oura, Fitbit, Apple Health, or labs sync happened, whether the coach prescribed today, why a workout or prescription didn't appear in the calendar, why wearable data looks stale or missing, whether an out-of-range lab marker needs a re-test, or wants to confirm health-mcp hook wiring after first-time setup.

The file

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 6 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Run health_lab_panel(today, lookback_days=180) and surface ANY marker with status: low or status: high:

    line 63
  2. Distinctive:

    If menstrual flow records exist:

    line 73
  3. Very specific:

    Read ~/.claude/settings.json to verify hooks are registered

    line 136
  4. Very specific:

    Read ~/.claude/hookify-blocks.log to verify hooks have fired recently

    line 137
  5. Very specific:

    Read <VAULT_ROOT>/Meta/coach-profile.yaml to verify profile state

    line 138
  6. Distinctive:

    PostToolUse on any health-mcp call: if Apple Health is > 28 days stale, surface a one-line nudge ("Re-export Apple Health — last import 31 …

    line 144
  7. Distinctive:

    SessionStart: if any lab is out-of-range AND > 90 days old, surface re-test reminder

    line 145
  8. Distinctive:

    Stop on coach-auto-prescribe: if today's prescription was created, surface the why_today line

    line 146
  9. Distinctive:

    No profile → report "/coach profile first"

    line 154
  10. Distinctive:

    No journal entries → skip cycle / Floor sections silently

    line 155

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

health-doctor

The substrate has 41 tools + 5 skills + 2 auto-trigger hooks. None of it matters if the chain breaks silently. This skill is the surface for verifying the system is actually running.

Run /health doctor whenever:

  • It's been a few days since you saw a workout in your calendar
  • You want to know if yesterday's wearable data actually made it in
  • You want to see which lab markers need a re-test
  • You're setting up the auto-chain for the first time and want to confirm the wiring

What it reports

Six sections, each with a green / yellow / red flag:

1. Data freshness

For each enabled source:

  • Apple Health: hours since last health_import_xml (yellow if > 14 days, red if > 28 days)
  • Oura: hours since last health_import_oura (yellow if > 36h, red if > 7d)
  • Fitbit: hours since last health_import_fitbit (yellow if > 36h, red if > 7d)
  • Labs: days since last health_import_labs (yellow if any marker is > 180 days old or any out-of-range marker is > 90d without a re-test)

Pulled from the imports table via health_status().

2. Last prescription + completion

  • Most recent prescription via health_coach_recent_prescriptions(days=7)
  • Was it completed? (RPE + lift actuals logged via /coach log)
  • Streak: how many days in a row had both a prescription AND a completion
  • Missed days: prescriptions with no completion in the last 14 days

Yellow if completion rate < 60%. Red if completion rate < 40% or no prescriptions in 7 days.

3. Auto-trigger hooks installed

The default wired set is defined by hooks.json at the repo root (the file the installer actually writes into ~/.claude/settings.json, tested by services/health-mcp/tests/test_v05_hooks.py) — read it rather than assuming a fixed list, so this check can't drift from what installs. See docs/AUTOMATION.md for the full chain.

By default that's one hook:

  • coach-auto-prescribe-on-journal.py on Stop (prescribes + backfills after /journal)

health-auto-sync.py (SessionStart wearable refresh) is opt-in, power-user only, and deliberately NOT wired by default — firing it on every SessionStart is wasteful for users with many sessions/day. Its absence from settings.json is expected, not a fault.

Red if the Stop hook (coach-auto-prescribe-on-journal.py) is not in settings.json. Yellow if the hook script file exists but isn't registered. Green if registered AND firing recently (check ~/.claude/hookify-blocks.log for the hook name in the last 48h). If health-auto-sync.py IS registered (the user opted in), report its freshness the same way, but never red for its absence.

4. Coach profile status

Reads <VAULT_ROOT>/Meta/coach-profile.yaml:

  • Profile exists? Last updated?
  • calendar_drop: true AND google-workspace MCP connected? (calendar drop won't fire otherwise)
  • preferred_workout_clock set?
  • days_per_week reasonable for level?
  • started_iso set? (used by deload-week computation)

Yellow if any field missing. Red if no profile at all.

5. Lab status flags

Run health_lab_panel(today, lookback_days=180) and surface ANY marker with status: low or status: high:

  • Marker, value, range, status
  • Days since last test for that marker
  • WHY this marker matters (pull from health_recommended_labs())
  • Suggested re-test cadence (e.g. "re-test in 90 days after supplementation")

Yellow if any marker is out-of-range and last tested > 90 days ago. Red if any out-of-range marker is critical (low ferritin in menstruating users, elevated hs-CRP > 3.0, fasting insulin > 10, Vitamin D < 20).

6. Cycle phase + sleep regularity (women's substrate qualifier)

If menstrual flow records exist:

  • Current phase + cycle day + irregularity flag from health_cycle_context(today)
  • Cycle length variance over last 6 cycles
  • Yellow if irregularity = "mild_irregular". Red if "irregular".

Plus health_sleep_regularity(last_14_days):

  • Regularity score
  • Bed-time stdev, wake-time stdev, mean sleep latency
  • Yellow if regularity < 70. Red if < 50.

Output format

A markdown report with the six sections, each summarized to 3-5 lines with the flag, the data, and a one-line "what to do" if yellow or red.

Example:

# Health doctor — 2026-05-10

## 🟢 Data freshness
- Apple Health: 12 days ago (yellow threshold: 14d) — re-export soon
- Oura: 4 hours ago — fresh
- Fitbit: not configured
- Labs: ApoB tested 2026-05-01 (9 days), Vitamin D last 2026-05-01 (9 days)

## 🟡 Last prescription + completion
- Last prescription: 2026-05-09 lower_body_strength (diff 7/10)
- Completed: no
- 7-day streak: 0 (missed yesterday's log)
- Action: /coach log yesterday's session to keep the progression chain accurate

## 🟢 Auto-trigger hooks
- SessionStart: health-auto-sync.py ✓ registered, last fired 4h ago
- Stop: coach-auto-prescribe-on-journal.py ✓ registered, last fired 14h ago

## 🟢 Coach profile
- /vault/Meta/coach-profile.yaml updated 2026-05-09
- calendar_drop: true (google-workspace MCP connected ✓)
- preferred_workout_clock: 07:00
- days_per_week: 4, level: intermediate, started_iso: 2026-05-09

## 🔴 Lab status flags
- Vitamin D 25-OH: 26 ng/mL (low, ref 30-100). Last tested 9 days ago.
  - Why: drives mood, immunity, recovery. Linked to chronic fatigue.
  - Suggested: 5000 IU/day; re-test in 90 days (target 2026-08-09)

## 🟡 Cycle + sleep regularity
- Cycle: luteal, day 22 (regular over last 6 cycles)
- Sleep regularity: 64/100 (yellow). Bed-time stdev 78min over last 14 days.
- Action: pick a wake time within a 30-min window for the next 14 days.

Tools called

  • health_status() — top-level table counts + imports table
  • health_coach_recent_prescriptions(days=7) — prescriptions + completion status
  • health_coach_summary(days=28) — completion rate
  • health_lab_panel(today, lookback_days=180) — most recent labs per marker
  • health_recommended_labs() — the WHY for any flagged marker
  • health_cycle_context(today) — current phase + irregularity
  • health_sleep_regularity(today-14, today) — bed/wake variance

Plus:

  • Read ~/.claude/settings.json to verify hooks are registered
  • Read ~/.claude/hookify-blocks.log to verify hooks have fired recently
  • Read <VAULT_ROOT>/Meta/coach-profile.yaml to verify profile state

When to surface unprompted

The hook system can surface specific flags WITHOUT the user running /health doctor explicitly:

  • PostToolUse on any health-mcp call: if Apple Health is > 28 days stale, surface a one-line nudge ("Re-export Apple Health — last import 31 days ago")
  • SessionStart: if any lab is out-of-range AND > 90 days old, surface re-test reminder
  • Stop on coach-auto-prescribe: if today's prescription was created, surface the why_today line

These are hookify nudges, configured separately. The skill is the comprehensive surface; the nudges are the targeted catches.

Graceful failure

  • health-mcp not registered → report "/health-setup first"
  • DuckDB empty → report "/ingest-health first"
  • No profile → report "/coach profile first"
  • No journal entries → skip cycle / Floor sections silently

The doctor never blocks. It always returns a report, even if it's "this is what's missing to get started."

Voice

Direct. Color-coded flags (🟢 / 🟡 / 🔴) for fast scan. Each yellow / red has a specific "what to do" line, not a vague "consider reviewing." Reader should know exactly what to fix in 30 seconds of reading.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.