healthclaw-guardrails

Survey ALL connected health data sources at once. Use when the patient asks: (1) "what's connected" or "what services are linked", (2) "check all my services / all my data", (3) "do you have my records" or "what records do you have", (4) "did my data come through from <Fasten/MEDENT/HealthEx/Health Bank One/Flexpa/Epic/wearables>". Calls fhir_get_token (for protected tenants) then sources_check, and presents connection status + record counts by type. Connection status and counts only — never clinical values.

HealthClaw Curatr (healthclaw.io) — patient-facing FHIR data quality evaluation and correction. Use when: (1) Evaluating a patient's health record for coding issues (deprecated code systems, invalid codes, missing required fields), (2) Presenting issues in plain language with clinical impact, (3) Applying patient-approved corrections with full Provenance tracking, (4) Preparing a structured correction request for the patient's healthcare provider. Supports FHIR R4 US Core v9 resources: Condition, AllergyIntolerance, MedicationRequest, Immunization, Procedure, DiagnosticReport — with ICD-10-CM, SNOMED CT, LOINC, CVX, and RxNorm validation via public terminology APIs.

Use this skill whenever connecting a patient's real health records from EHR systems (Epic, Cerner, Athena) or the TEFCA national network into HealthClaw Guardrails. Covers: Fasten Stitch widget embed, org_connection_id registration, EHI export job tracking, NDJSON ingestion status, TEFCA IAS identity-verified multi-provider retrieval, and post-import Curatr quality scan workflow.

HealthClaw Guardrails (healthclaw.io) — FHIR agent guardrails for clinical data access via MCP. Supports FHIR R4 US Core v9 (stable) and FHIR R6 ballot3 (experimental). Use when: (1) Reading patient data through MCP tools with automatic PHI redaction, (2) Writing clinical resources with two-phase propose/commit and step-up authorization, (3) Querying observation statistics or recent lab results, (4) Evaluating R6 Permission resources for access control decisions, (5) Auditing agent access to healthcare data. 14 MCP tools.

Connect to real FHIR servers through the MCP guardrail proxy. Use when: (1) Connecting to HAPI FHIR, SMART Health IT, or Epic sandbox servers, (2) Proxying AI agent requests to production EHR systems with guardrails, (3) Ensuring upstream server URLs never leak to clients, (4) Understanding how redaction, audit, and step-up auth apply to upstream data.

End-to-end onboarding for the HealthClaw + OpenClaw personal-health-agent stack. Walks anyone — not just developers — through (1) installing OpenClaw as the local AI gateway, (2) standing up an open-source FHIR server (HAPI or Medplum) on their machine, (3) connecting their EHR records via HealthEx, Josh Mandel's fhir-skills, Flexpa, or any patient-right-of-access / TEFCA IAS service, (4) installing HealthClaw Guardrails with the OpenClaw personas pre-wired, and (5) pulling + reviewing health data end-to-end. Triggers when a user asks "how do I get started", "how do I set this up", "what do I need to use HealthClaw", "first time setup", "install everything", or any first-run question. Use this before any other HealthClaw skill when the user hasn't yet built their stack.

Pull verified medical records and digital-identity context from Health Bank One (https://www.healthbankone.com) via their MCP server. HBO uses OAuth 2.x with per-consumer authorization, so this skill drives an authorization-code grant first, then runs the MCP pull, then ingests the redacted bundle into a HealthClaw tenant. Use when the user wants to add HBO as a source alongside HealthEx and Fasten Connect. Triggers on prompts like "connect Health Bank One", "pull from HBO", "verify identity through Health Bank One".

HealthClaw HealthEx Export (healthclaw.io) — automated personal health record export from the local HealthClaw FHIR store. Use when: (1) The patient wants to export all their health data from the HealthClaw local store as a portable FHIR bundle, (2) Migrating health data to a new tenant or archive, (3) Creating a de-identified snapshot for sharing with a provider or second opinion, (4) Pre-screening records for Curatr quality issues before a full evaluation, (5) Automating the HealthEx → local FHIR store ingestion pipeline.

HealthEx → HealthClaw-redacted export via the official MCP Python SDK (mcp>=1.2). Use when: (1) Pulling fresh clinical data from HealthEx as the upstream source of truth (not from the local HealthClaw store), (2) Writing a PHI-redacted snapshot to disk before any ingest, so the raw MCP response never hits the filesystem, (3) Producing a single-file JSON or NDJSON bundle for downstream import via `/import`, (4) Running the pipeline headlessly from a Telegram bot or cron on the Mac mini. For the older direct-REST pull against the local FHIR store, see the `healthex-export` skill.

Run the HealthClaw guardrail stack inside Hermes, Nous Research's self-improving AI agent. Same conversational gateway and same compliance goals as the OpenClaw integration, but with Hermes' learning loop — skills that improve from experience, conversation memory across sessions, and native MCP support over Streamable HTTP. Triggers when a user asks about Hermes, the agentskills.io standard, SOUL personas, the Nous agent, or wants an alternative to OpenClaw that learns over time.

Sally-PCP hypertension follow-up coordination skill. Use when: (1) Running BP check-in calls for landline patients (Rosa pattern — Bland.ai outbound voice with staff confirmation), (2) Flagging undiagnosed elevated readings (≥3 observations ≥140/90 across encounters) for PCP discussion, (3) Escalation scheduling when readings trend above the patient's baseline, (4) Medication refill coordination via practice pharmacy line, (5) Smartphone patient coordination via Telegram or SMS nurse-line texts (Marcus pattern). Covers the full propose → patient/staff confirms → fhir_get_token → action_commit loop using MCP action tools. Administrative coordination only — never clinical advice, never medication adjustments.

Connect your health records from any US health system via HealthEx, pull your complete clinical history, analyze it with Claude, and optionally export to a personal de-identified FHIR store with automated data quality curation via HealthClaw Guardrails. Supports Epic, Cerner, CommonWell, Carequality, and most major US EHR networks. Triggers when a user asks to connect health records, pull medical history, review lab results, check immunizations, identify care gaps, or export data to a personal FHIR store.

Conservative PHI redaction patterns for FHIR resources. Use when: (1) Redacting patient health information from FHIR resources before AI agent access, (2) Implementing de-identification for clinical data pipelines, (3) Understanding what fields are stripped, masked, or truncated in FHIR resources, (4) Building read paths that automatically protect patient privacy.

Generate a patient-controlled SMART Health Link (SHL) QR code for sharing health records with a clinic or provider. Use when: (1) A patient asks to share their records via QR for clinic check-in, (2) A patient asks to generate a health QR code, (3) A patient asks to "share my records" or "generate a QR with my records". Covers the full consent → fhir_get_token → shl_generate → QR render + manage-link delivery loop. Never direct-encodes PHI into a QR image — the QR encodes only the encrypted shlink:/ pointer.