agent-team-spec · supabase_json

Limitless OS AI-First Agent Team Specification

agent-team-spec artifact · for Limitless OS · status draft

Agent Org Chart

Limitless OS runs like an AI-first operating team.

Each pod has a job, a queue, a memory scope, and a human handoff rule. The chart shows who coordinates work, who operates the desk, who supports care, who grows the business, and where guardrails sit.

Pods

operating groups

Agents

defined roles

Human Owner Seth / Limitless Team Approves strategy, clinical judgment, patient-sensitive actions, and external commitments.

Operating Brain Chief Operating Brain Owns daily priority synthesis, cross-agent routing, owner focus, and operating decisions.

15Phase 1

16Phase 2

8Phase 3

executive_cognition

Executive Agents

Think across the business, synthesize signal, prioritize work, and preserve decisions.

Chief Operating Brain Phase 1

Owns daily priority synthesis, cross-agent routing, owner focus, and operating decisions.

Strategy / Growth Advisor Phase 2

Turns business context, offers, referral channels, and market opportunities into growth options.

Daily Briefing Agent Phase 1

Creates the morning operating brief: today, risks, opportunities, and decisions needed.

Decision Logger Phase 1

Captures decisions, rationale, owner, date, and follow-up actions from chats, notes, and meetings.

Scoreboard Analyst Phase 1

Interprets lead, schedule, conversion, revenue, retention, outcome, and queue-health signals.

front_office_action

Front Office Agents

Operate the business desk: intake, scheduling, follow-up, referrals, and admin triage.

Intake Concierge Phase 1

Owns first-response intelligence for new leads, missed calls, web forms, DMs, and referrals.

Scheduler Phase 1

Owns booking flow, schedule gaps, cancellation recovery, waitlist fills, and follow-up timing.

Missed Call Recovery Agent Phase 1

Detects missed calls and creates rapid response drafts with lead/patient context.

Cancellation Recovery Agent Phase 1

Turns cancellations and no-shows into recovery actions, waitlist fills, and risk alerts.

Patient Communication Agent Phase 1

Drafts empathetic patient texts/emails for scheduling, reminders, follow-up, and education handoffs.

Referral Follow-up Agent Phase 2

Maintains physicians, coaches, gyms, schools, trainers, and partner relationship follow-ups.

Insurance / Admin Triage Agent Phase 2

Flags missing admin information, payer/authorization issues, balances, and payment friction for review.

clinical_reasoning_support

Clinical Support Agents

Support care continuity while keeping clinician judgment in control.

Visit Prep Agent Phase 2

Prepares clinicians before visits with last context, goals, risk flags, and items to address.

Plan of Care Companion Phase 2

Tracks continuity against the plan of care and surfaces adherence, milestone, and communication needs.

Documentation Drafter Phase 2

Turns visit notes, voice notes, and observations into structured documentation drafts.

Patient Progress Monitor Phase 2

Looks for adherence risk, plateau, symptom changes, or progress signals between visits.

Home Program Follow-up Agent Phase 3

Supports adherence with reminders, encouragement, and clinician-approved education.

Clinical Question Triage Agent Phase 3

Classifies incoming clinical questions by urgency and prepares clinician review packets.

performance_intelligence

Performance / Athlete Agents

Use testing and return-to-sport context to make progress visible and actionable.

Return-to-Sport Agent Phase 3

Summarizes readiness milestones, gaps, confidence, and clinician review needs for athletes.

Testing Data Interpreter Phase 3

Translates VALD/ForceDecks/Dynamo data into understandable summaries and review packets.

Strength / Force Trend Agent Phase 3

Detects improvements, regressions, asymmetries, and plateau patterns in strength/force data.

Runner / Athlete Progress Agent Phase 3

Tracks athlete-specific progress narratives, confidence, training tolerance, and next discussion points.

Readiness Check-in Agent Phase 3

Collects and summarizes soreness, confidence, sleep/recovery, adherence, and readiness check-ins.

Performance Report Agent Phase 3

Creates patient, parent, coach, and internal performance report drafts from approved data.

growth_voice

Growth / Voice Agents

Extend Seth’s expertise and brand voice into content, education, referral nurture, and reactivation.

Content Brain Phase 2

Turns Seth’s captured thoughts, clinical philosophy, patient themes, and local authority into content strategy.

Social Post Drafter Phase 2

Creates post drafts, hooks, reels scripts, captions, and repurposing packages.

Newsletter Agent Phase 2

Builds educational, relationship-building, and campaign newsletter drafts.

Patient Education Agent Phase 2

Turns clinical expertise into plain-language, patient-safe educational explanations.

Referral Partner Nurture Agent Phase 2

Creates partner touchpoints, updates, and collaboration ideas for referral relationships.

Testimonial / Review Opportunity Agent Phase 2

Identifies appropriate review/testimonial opportunities and drafts compliant requests.

memory_formation

Memory / Knowledge Agents

Make the operating system smarter over time by capturing, curating, and reusing knowledge.

Thought Capture Agent Phase 1

Captures Seth’s brainstorming from ChatGPT/Claude and stores usable operating memory.

Operating Memory Curator Phase 1

Turns raw thoughts into structured principles, workflows, tags, and reusable context.

Patient Context Summarizer Phase 2

Creates concise context packets from authorized patient references for clinician/front-office review.

Protocol Librarian Phase 2

Organizes SOPs, clinical education snippets, scripts, and reusable office protocols.

Workflow Improvement Agent Phase 2

Finds repeated friction and proposes process changes, automations, or SOP updates.

governance

Governance Agents

Protect privacy, clinical safety, patient trust, and automation boundaries across every pod.

HIPAA / Privacy Guardrail Phase 1

Checks PHI boundaries, HCS payload safety, patient data exposure, and connector/tool access.

Clinical Safety Reviewer Phase 1

Reviews content for diagnosis, treatment, return-to-play, red flag, and individualized advice risk.

Automation Boundary Agent Phase 1

Decides whether an action may be automated, drafted, or must be human-only.

Human Approval Router Phase 1

Routes drafts and decisions to Seth, Betsy/front office, clinician, or performance team based on sensitivity.

Build Sequence

Roll this out as an operating team, not a feature dump.

Thought-to-Action Operating Loop

Give Seth a real AI-first loop: brainstorm -> capture -> prioritize -> queue -> draft action -> approve.

Memory / Knowledge AgentsExecutive AgentsFront Office AgentsGovernance Agents

Growth and Office Expansion

Add brand voice, referral nurture, admin triage, visit prep, documentation drafting, and workflow improvement once the V1 loop is trusted.

Growth / Voice AgentsFront Office AgentsClinical Support AgentsMemory / Knowledge Agents

Clinical and Performance Intelligence

Add deeper clinician-reviewed patient progress, home program, clinical triage, and performance testing intelligence.

Clinical Support AgentsPerformance / Athlete Agents

Planning Surface

Use this to decide what happens next.

Status

draft

Agent Handoff

Start Here

agent-team-spec artifact · for Limitless OS · status draft

Completion Evidence

No explicit evidence field yet. Require tests, screenshots, linked PRs, or reviewed outputs before marking complete.

Artifact Shape

pods: 7 items
title: Limitless OS AI-First Agent Team Specification
source: atlas-codex
client id: limitless
hcs model: principle: string, review_modes: object, queue_item_required_fields: object
project id: 07dee66f-8143-4f4d-9af9-b1ed413d2767
updated at: 2026-05-09T19:51:33.269Z
project name: Limitless OS
artifact type: agent-team-spec
agent contract: principle: string, required_fields: object, evaluation_question: string
build sequence: 3 items
schema version: 2

Structured Payload

Machine-readable source fields

pods

id	title	agents	purpose
executive_cognition	Executive Agents	- Chief Operating Brain: id: string, name: string, role: string, phase: number, queue: string, tool_set: object, memory_scope: object, confidence_model: string, human_handoff_rule: string - Strategy / Growth Advisor: id: string, name: string, role: string, phase: number, queue: string, tool_set: object, memory_scope: object, confidence_model: string, human_handoff_rule: string - Daily Briefing Agent: id: string, name: string, role: string, phase: number, queue: string, tool_set: object, memory_scope: object, confidence_model: string, human_handoff_rule: string - Decision Logger: id: string, na	Think across the business, synthesize signal, prioritize work, and preserve decisions.
front_office_action	Front Office Agents	- Intake Concierge: id: string, name: string, role: string, phase: number, queue: string, tool_set: object, memory_scope: object, confidence_model: string, human_handoff_rule: string - Scheduler: id: string, name: string, role: string, phase: number, queue: string, tool_set: object, memory_scope: object, confidence_model: string, human_handoff_rule: string - Missed Call Recovery Agent: id: string, name: string, role: string, phase: number, queue: string, tool_set: object, memory_scope: object, confidence_model: string, human_handoff_rule: string - Cancellation Recovery Agent: id: string, name:	Operate the business desk: intake, scheduling, follow-up, referrals, and admin triage.
clinical_reasoning_support	Clinical Support Agents	- Visit Prep Agent: id: string, name: string, role: string, phase: number, queue: string, tool_set: object, memory_scope: object, confidence_model: string, human_handoff_rule: string - Plan of Care Companion: id: string, name: string, role: string, phase: number, queue: string, tool_set: object, memory_scope: object, confidence_model: string, human_handoff_rule: string - Documentation Drafter: id: string, name: string, role: string, phase: number, queue: string, tool_set: object, memory_scope: object, confidence_model: string, human_handoff_rule: string - Patient Progress Monitor: id: string,	Support care continuity while keeping clinician judgment in control.
performance_intelligence	Performance / Athlete Agents	- Return-to-Sport Agent: id: string, name: string, role: string, phase: number, queue: string, tool_set: object, memory_scope: object, confidence_model: string, human_handoff_rule: string - Testing Data Interpreter: id: string, name: string, role: string, phase: number, queue: string, tool_set: object, memory_scope: object, confidence_model: string, human_handoff_rule: string - Strength / Force Trend Agent: id: string, name: string, role: string, phase: number, queue: string, tool_set: object, memory_scope: object, confidence_model: string, human_handoff_rule: string - Runner / Athlete Progres	Use testing and return-to-sport context to make progress visible and actionable.
growth_voice	Growth / Voice Agents	- Content Brain: id: string, name: string, role: string, phase: number, queue: string, tool_set: object, memory_scope: object, confidence_model: string, human_handoff_rule: string - Social Post Drafter: id: string, name: string, role: string, phase: number, queue: string, tool_set: object, memory_scope: object, confidence_model: string, human_handoff_rule: string - Newsletter Agent: id: string, name: string, role: string, phase: number, queue: string, tool_set: object, memory_scope: object, confidence_model: string, human_handoff_rule: string - Patient Education Agent: id: string, name: string	Extend Seth’s expertise and brand voice into content, education, referral nurture, and reactivation.
memory_formation	Memory / Knowledge Agents	- Thought Capture Agent: id: string, name: string, role: string, phase: number, queue: string, tool_set: object, memory_scope: object, confidence_model: string, human_handoff_rule: string - Operating Memory Curator: id: string, name: string, role: string, phase: number, queue: string, tool_set: object, memory_scope: object, confidence_model: string, human_handoff_rule: string - Patient Context Summarizer: id: string, name: string, role: string, phase: number, queue: string, tool_set: object, memory_scope: object, confidence_model: string, human_handoff_rule: string - Protocol Librarian: id: st	Make the operating system smarter over time by capturing, curating, and reusing knowledge.
governance	Governance Agents	- HIPAA / Privacy Guardrail: id: string, name: string, role: string, phase: number, queue: string, tool_set: object, memory_scope: object, confidence_model: string, human_handoff_rule: string - Clinical Safety Reviewer: id: string, name: string, role: string, phase: number, queue: string, tool_set: object, memory_scope: object, confidence_model: string, human_handoff_rule: string - Automation Boundary Agent: id: string, name: string, role: string, phase: number, queue: string, tool_set: object, memory_scope: object, confidence_model: string, human_handoff_rule: string - Human Approval Router:	Protect privacy, clinical safety, patient trust, and automation boundaries across every pod.

title

Limitless OS AI-First Agent Team Specification

source

atlas-codex

client id

limitless

hcs model

principle

Every agent should create a queue item, pull a queue item, review/escalate a queue item, or summarize queue health.

review modes

auto_log_only
draft_requires_approval
human_must_act
blocked_needs_policy_decision

queue item required fields

task_id
role_topic_slug
source_refs
sensitivity
patient_data_boundary
draft_output_refs
approval_required
confidence
next_human_owner

project id

07dee66f-8143-4f4d-9af9-b1ed413d2767

updated at

2026-05-09T19:51:33.269Z

project name

Limitless OS

artifact type

agent-team-spec

agent contract

principle

Agents are AI-native collaborators, not software modules with agent names.

required fields

role
memory_scope
tool_set
queue
confidence_model
human_handoff_rule

evaluation question

Can this agent make judgment, remember context, pull work, use tools, express confidence, and hand off safely?

build sequence

name	phase	purpose	included pods	first workflows	included agents
Thought-to-Action Operating Loop	V1	Give Seth a real AI-first loop: brainstorm -> capture -> prioritize -> queue -> draft action -> approve.	- memory_formation - executive_cognition - front_office_action - governance	- ChatGPT/Claude thought capture - daily operating brief - front-office intake queue - missed call recovery - cancellation recovery - draft patient communication with approval	- chief_operating_brain - daily_briefing_agent - decision_logger - scoreboard_analyst - intake_concierge - scheduler - missed_call_recovery - cancellation_recovery - patient_communication - thought_capture_agent - operating_memory_curator - hipaa_privacy_guardrail - clinical_safety_reviewer - automation_boundary_agent - human_approval_router
Growth and Office Expansion	V2	Add brand voice, referral nurture, admin triage, visit prep, documentation drafting, and workflow improvement once the V1 loop is trusted.	- growth_voice - front_office_action - clinical_reasoning_support - memory_formation	- content from captured thoughts - referral follow-up - admin issue triage - upcoming visit prep - documentation draft - SOP/protocol library	- strategy_growth_advisor - referral_followup - insurance_admin_triage - visit_prep_agent - plan_of_care_companion - documentation_drafter - patient_progress_monitor - content_brain - social_post_drafter - newsletter_agent - patient_education_agent - referral_partner_nurture - testimonial_review_opportunity - patient_context_summarizer - protocol_librarian - workflow_improvement_agent
Clinical and Performance Intelligence	V3	Add deeper clinician-reviewed patient progress, home program, clinical triage, and performance testing intelligence.	- clinical_reasoning_support - performance_intelligence	- patient progress flags - home program follow-up - clinical question triage - VALD/ForceDecks summary - return-to-sport review packet	- home_program_followup - clinical_question_triage - return_to_sport_agent - testing_data_interpreter - strength_force_trend_agent - runner_athlete_progress_agent - readiness_checkin_agent - performance_report_agent

schema version

source documents

path	role	pages	title	date
/Users/justinking/Downloads/Limitless_OS_Final_Vision_Document.pdf	Patient/performance operating system vision	19	Limitless OS Final Vision & Goals/Intent Document
/Users/justinking/Downloads/Limitless Ai Agent Goals Intent Blueprint.pdf	Front-office and business-support agent blueprint	40	Limitless Physical Therapy & Performance AI Agent Goals & Intent Blueprint
	AI-first agent grouping and MCP thought capture design		Atlas-Codex working session	2026-05-09

design principles

Always think agents AI-first: role, judgment, memory, tools, queue, confidence, and human handoff come before implementation details.
Pods own workflows; agents own jobs; HCS queues own handoff and auditability.
Agents may synthesize, draft, recommend, route, and summarize. They only act autonomously where a policy explicitly allows it.
Patient-facing and clinical outputs are draft/review-first until privacy, clinical safety, and approval rules are proven.
PT Everywhere remains the operational backbone; Limitless OS wraps it with memory, queues, orchestration, and decision support.

v1 recommendation

name

Front-office revenue flow plus shared command layer

rationale

This is operational, measurable, lower clinical risk, and gives Seth immediate leverage while the clinical/performance layers mature behind guardrails.

included agents

chief_operating_brain
daily_briefing_agent
decision_logger
scoreboard_analyst
intake_concierge
scheduler
missed_call_recovery
cancellation_recovery
patient_communication
thought_capture_agent
operating_memory_curator
hipaa_privacy_guardrail
clinical_safety_reviewer
automation_boundary_agent
human_approval_router

success metrics

connector captures real Seth thoughts
daily briefing produces useful priorities
new lead response time improves
missed calls become tracked follow-ups
cancellations produce recovery actions
human approval queue is clear and trusted