Mirobody is a health data engine that collects information from labs, wearables, genomics, imaging, and other sources, then standardizes it for AI use. It supports health applications and agents that answer questions about a person’s medical data with charts and citations. The catalogue skill extends interaction with this engine.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add thetahealth/mirobody --skill lab-report-walkthroughgit clone --depth 1 https://github.com/thetahealth/mirobodyWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/thetahealth/mirobody/lab-report-walkthrough)<a href="https://agentmods.dev/skills/thetahealth/mirobody/lab-report-walkthrough"><img src="https://agentmods.dev/badge/skills/thetahealth/mirobody/lab-report-walkthrough/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/thetahealth/mirobody/lab-report-walkthrough"><img src="https://agentmods.dev/badge/skills/thetahealth/mirobody/lab-report-walkthrough.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00067 | $0.00848 |
| Opus 5 | $0.00034 | $0.00424 |
| Sonnet 5 | $0.00013 | $0.00170 |
| Haiku 4.5 | $0.00007 | $0.00085 |
Grade A, and why
lab-report-walkthrough scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 69 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Lab Report Walkthrough
Turn a raw lab report into an explanation a person can act on, without ever drifting into diagnosis.
Workflow
-
Read the original, not a summary. The report is under
/uploads/(this conversation) or/library/(history).read_filethe PDF/image itself — layout carries meaning: grouped panels, flags (H/L/↑/↓), footnotes about specimen or method. -
Organize by panel, not by document order. Group results the way clinicians think: lipids, CBC, metabolic, liver, kidney, thyroid, vitamins, urinalysis. Keep the original indicator names — do not translate them — and keep each value with its unit and printed reference range.
-
Flag against the PRINTED reference range. Ranges differ by lab, method, age and sex; the range on the page beats any range you remember. Sort the walkthrough: out-of-range first, borderline second, normal last (summarize normals in one line — nobody needs fifteen "this is fine" paragraphs).
-
Compare with history when it exists. A single value is a dot; two are a direction. Check
/library/for prior reports of the same panel. A value inside range but moving steadily toward a boundary is worth a sentence; a chart is better — emit avis-chartblock for any indicator with 3+ points over time. -
Explain in plain language, three sentences per flagged item: what the indicator measures, what direction it moved and by how much, and what commonly influences it (fasting state, hydration, recent exercise, common medications). Name the influence factors as possibilities, not conclusions.
Boundaries
- No diagnosis, no treatment advice. "Your ALT is 2× the upper limit" is a fact; "you have liver disease" is a diagnosis. State facts, then recommend discussing flagged results with their clinician — specifically, by name: "worth asking your doctor about the ALT and AST together".
- Never invent a value or a range. If the scan is illegible for a row, say so and list what you could not read.
- Keep the original language of the report for indicator names, adding a translation in parentheses when the conversation is in another language.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 69 lines · 67 tokens per session scan A ee18d96f1802
lab-report-walkthrough is a skill published in the GitHub repository thetahealth/mirobody (1,296 stars, last pushed yesterday), licensed Apache-2.0. It adds 67 tokens to every session and 848 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
hypertension-coordinator
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…
personal-health-records
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…
healthbankone
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…
curatr
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…
fasten-connect
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, orgconnectionid registration, EHI export job tracking, NDJSON ingestion status, TEFCA IAS identity-verified…
hermes
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…