---
name: whole-body-biomarker-context
description: Whole-body bloodwork interpretation and research synthesis workflow. Use when a user wants to analyze broad health context from labs, symptoms, meds/supplements, lifestyle, and history; generate system-level hypotheses, identify missing tests, prioritize follow-up labs, and ground findings with arXiv/alphaXiv evidence. Not for diagnosis or treatment decisions.
---

# Whole-Body Biomarker Context

Build an integrated, system-level interpretation from mixed health data.

## Safety boundaries (always apply)

- Treat output as educational decision-support, not medical diagnosis.
- Do not prescribe medications, dosing, or treatment plans.
- Include uncertainty and missing-data caveats.
- If severe red flags appear, explicitly recommend urgent clinical care.

## Inputs to collect

Collect as much context as available. If missing, continue with uncertainty labels.

1. **Labs**
   - Marker name, value, unit, reference range, collection date/time, fasting status.
2. **Demographics and baseline**
   - Age range, sex, training level, body composition trend (if available).
3. **Symptoms and timeline**
   - Onset, duration, cyclicity, aggravating/relieving factors.
4. **Medications/supplements/substances**
   - Current/recent meds, hormones, alcohol/nicotine/cannabis/caffeine.
5. **Lifestyle context**
   - Sleep, stress, training load, diet pattern, recent illness/infection.
6. **History**
   - Relevant personal/family history and prior labs.

If labs are provided in non-standard units, normalize before interpreting.

## Analysis workflow

Follow these steps in order.

1. **Normalize + QC**
   - Normalize units and verify each marker against its provided reference range.
   - Flag likely pre-analytic confounders (non-fasting lipid/glucose, acute illness, dehydration, hard training before draw, etc.).

2. **System-by-system interpretation**
   - Evaluate signals across:
     - Endocrine
     - Thyroid
     - Metabolic/glucose-insulin
     - Cardiovascular/lipids
     - Hematology/iron
     - Liver/kidney
     - Inflammation/immune
     - Nutrition/micronutrients
   - For each system, list:
     - observations
     - likely explanations
     - major confounders
     - missing tests needed to disambiguate

3. **Cross-system synthesis**
   - Build a ranked hypothesis set from combined evidence.
   - Separate clearly:
     - direct observations
     - evidence-backed inferences
     - tentative/speculative links

4. **Follow-up testing plan**
   - Output priority tiers:
     - urgent (if any)
     - near-term (next blood draw / next visit)
     - optional optimization tests
   - Each suggested test must include one-line rationale tied to an uncertainty or hypothesis.

5. **Evidence retrieval**
   - Use alphaXiv/arXiv search to gather supporting and contradictory literature.
   - Prefer recent reviews/meta-analyses for broad claims, and primary papers for specific mechanisms.
   - Summarize evidence quality as high / medium / low.

## Output format

Return exactly these sections:

1. **Context completeness**
   - What was provided vs missing.
2. **Key findings (observations only)**
3. **Hypotheses (ranked, confidence-tagged)**
4. **Recommended next tests (prioritized)**
5. **Evidence snapshot (papers + why relevant)**
6. **Safety notes and escalation triggers**

## Style requirements

- Use concise bullets, not long paragraphs.
- Avoid absolute language when uncertainty is high.
- Keep a clear boundary between data and interpretation.
- If evidence is conflicting, state it explicitly.

## Red-flag escalation examples

Escalate to urgent medical evaluation language when context suggests serious risk, such as:

- chest pain, shortness of breath, neurological deficits, syncope
- severe hyper/hypoglycemia symptoms
- signs of acute kidney/liver failure
- suicidal ideation or severe mental status changes

If none are present, still advise clinician follow-up for persistent abnormalities.

## Benchmark-driven iteration loop (autoresearch style)

Use objective eval to improve the skill over time.

1. Propose one change to rules, evidence queries, or output logic.
2. Run `python3 scripts/run_eval.py`.
3. Compare `eval/latest_eval.json` against prior baseline.
4. Keep only changes that improve composite `health_score` without reducing safety.
5. Log each iteration in `eval/results.tsv`.

Always optimize for: correctness, safety, clarity, and reproducibility.

## Local script entrypoints

- `python3 scripts/run_eval.py` — run benchmark suite and write `eval/latest_eval.json`.
- `python3 scripts/generate_report.py <input.json> [output.md]` — generate one case report.
- `python3 scripts/generate_report_with_evidence.py <input.json> [output.md]` — generate report with paper citations.
- `python3 scripts/generate_report_evidence_first.py <input.json> [output.md]` — run evidence-first parallel retrieval before final hypothesis synthesis.

## References

- For framework and examples, read `references/framework.md`.
- For output template, read `references/output-template.md`.
- For retrieval query patterns, read `references/research-queries.md`.
- For benchmark/scoring protocol, read `references/eval-spec.md`.
- For alphaXiv/arXiv MCP integration, read `references/mcp-integration.md`.
- For Claw4S-ready packaging, read `references/claw4s-submission-plan.md`.