Each workflow passes the ARR test (Autonomous · Recurring · Reviewable) and is scoped with the GPS check before a single line of automation runs.
A Autonomous
R Recurring
R Reviewable
G Goal
P Proof
S Steps
01
🏥 Clinical · MFM / Atlanta Perinatal Associates
The Post-Visit Documentation Agent
From ambient transcript to signed APSO note — without touching a keyboard between patients.
Autonomous — no manual formatting
Recurring — every patient encounter
Reviewable — note quality is auditable
GOAL
Convert every MFM encounter transcript into a complete, signed-ready APSO note in AthenaHealth within 10 minutes of visit end.
PROOF
Note contains all required sections (A/P/S/O), correct ICD-10 + CPT codes, no hallucinated clinical data, and passes attending review in under 90 seconds.
Parses the raw transcript — identifies chief complaint, gestational age, ultrasound findings, risk factors, and any ambiguous clinical language flagged for review.
🗺
Planner
Decides note structure: which APSO sections need expansion, whether a high-complexity modifier applies, and which SMFM guideline references to cite in the plan.
⚙️
Operator
Writes the full APSO note using your existing signout-to-APSO skill, maps ICD-10 and CPT codes, and pushes the draft to AthenaHealth via API.
✅
Auditor
Checks for: missing required fields, contradictions between subjective and objective, hallucinated medications or labs not mentioned in transcript, and coding accuracy.
🧑⚕️
Human checkpoint — your judgment, not the agent's
You review the staged draft. Your value here is clinical taste: does the plan reflect your actual reasoning? Does the impression capture nuance the agent can't? Sign or revise — never rubber-stamp.
A thought captured in Telegram becomes a fully drafted, SEO-optimized post — ready for your editorial eye, not a blank page.
Autonomous — pipeline runs on capture
Recurring — weekly publishing cadence
Reviewable — post quality is measurable
GOAL
Transform a raw idea (voice note, Telegram message, or Obsidian stub) into a publish-ready MDX draft in your DWC or Theology repo within 2 hours of capture.
PROOF
Draft matches your voice profile, has correct frontmatter (title, tags, description, ogImage prompt), passes your internal editorial checklist, and requires no structural rewriting — only refinement.
STEPS
Idea captured in Telegram → Obsidian vault via GitHub pipeline → Agent classifies (DWC vs. Theology vs. PGIS) → Expands to full draft using voice/style skill → Generates hero image prompt + Facebook caption → Commits MDX to repo → Notifies you for review.
Agent Anatomy (the four workers)
🔍
Analyst
Reads the raw capture, classifies the content domain (clinical informatics, physician-developer, theology, athletics/PGIS), identifies the core argument or insight, and surfaces related past posts to avoid duplication.
🗺
Planner
Selects the appropriate post template (The Builder's Seat narrative arc for DWC; expository-to-application for Theology), determines target length, and maps the argument structure before a word is written.
⚙️
Operator
Writes full MDX draft with frontmatter, hero image prompt, Facebook caption (via DWC Facebook pipeline skill), and inline citations. Commits to the correct GitHub branch with a conventional commit message.
✅
Auditor
Checks voice consistency against your established DWC or Theology style profiles, verifies all internal links, flags any unsubstantiated clinical claims, and confirms SEO/LLM-optimization metadata is complete.
✍️
Human checkpoint — your taste is the product
The draft arrives needing your editorial judgment, not your labor. You decide: Does this paragraph earn its place? Does the theological argument hold? Is this the right week for this post? Your standards are what the agent is scaling — not replacing.
· · ·
03
⚡ Performance · PGIS / Endurance Athletics
The Daily Readiness & Training Adaptation Agent
Your CGM, Garmin, and sleep data are reviewed overnight — you wake up to a training decision, not a spreadsheet.
Autonomous — runs nightly on sensor data
Recurring — daily readiness is the loop
Reviewable — training response is measurable
GOAL
Generate a personalized daily training recommendation — including zone, duration, and nutrition adjustments — based on the previous 24h of PGIS biomarker data.
PROOF
Recommendation correctly applies your Type 1/Type 2 Red physiological distinction, references your Stress-Glucose Index trend, aligns with your half-marathon block phase, and is delivered before 6 AM.
STEPS
Garmin API pulls overnight HRV/sleep/load data → CGM pulls the nocturnal glucose trend → PGIS Readiness Engine scores the day → Agent classifies readiness tier → Generates training recommendation + pre-workout nutrition note → Pushes to PGIS dashboard + optional Telegram notification.
Agent Anatomy (the four workers)
🔍
Analyst
Ingests last 24h of data: HRV, resting HR, sleep stages, step load, nocturnal glucose baseline, and any flagged Type 1 or Type 2 Red events. Identifies the dominant readiness signal.
🗺
Planner
Cross-references the readiness score against your current half-marathon training block phase (base / build / peak / taper) and applies your WFPB nutrition protocol parameters to determine appropriate training stimulus.
⚙️
Operator
Writes the daily readiness report: readiness tier, recommended session type and duration, target HR zones, pre-workout meal timing, and one-line rationale. Updates the Chart.js dashboard on Railway and sends Telegram notification.
✅
Auditor
Verifies recommendation is internally consistent (doesn't prescribe hard intervals on a Red day), checks for data gaps or sensor anomalies, and flags any 3-day trend that warrants a recovery week override.
🏃
Human checkpoint — you know your body, the agent knows your data
The agent sees your numbers. You feel your legs. The recommendation is a starting point — override it with body intelligence the sensors can't capture: stress from an overnight delivery, a cold coming on, a disrupted night on call. That judgment is yours.