agentleFS
shreyasugemugeshreyasugemuge/claude-health-coach/.claude/skills/ingest/SKILL.md

ingest

Skill0 starsChanged 6 months agoMITView on GitHub

Process any file dropped into raw/. Bloodwork PDFs (parse values, normalize units, compare to India thresholds, update biomarkers, generate biomarker-event review). Medical PDFs (extract diagnoses, update profile/medical). Wearable exports (Apple Health zip, Oura csv, Apple Watch). Grocery receipts from Zepto/BigBasket/Blinkit (OCR, update inventory, flag restock items). Body composition photos (file under raw/photos/body/, update biomarkers/body-composition). Random notes from raw/inbox/ (triage to right destination). Invoke when the user drops a file, says "ingest X", "I uploaded my bloodwork", "here's my report", "I dropped my Oura export", "process inbox", "parse this PDF", or when a file appears in raw/ that hasn't been processed yet (check log.md for last ingest event of each filename).

90 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use daily-arxiv, exp-eval and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • LambenthanSkill

    daily-arxiv

    每日从 arXiv 拉取新论文,过滤相关性,auto-ingest 高优先级论文,检测 SOTA 更新

    Also in 5 projects that use this

  • LambenthanSkill

    exp-eval

    实验判决门:Review LLM 独立评判实验结果 → 4 种判决路径 → 自动更新 claims confidence、ideas status、graph edges

    Also in 5 projects that use this

  • phanghonghaoSkill

    poster

    HTML/CSS A4 海报生成。从沙龙/活动 Markdown 文件生成科技风 A4 海报 PDF/PNG。Use when user mentions: poster, 海报, 海报生成, make poster, generate poster, 活动海报, 沙龙海报.

    Also in 5 projects that use this

  • skyllwtSkill

    exp-pilot-eval

    Pilot result evaluation — read pilot results, apply success criteria, update idea page (pilot_result, failure_reason if failed), generate PILOT_VERDICT_REPORT. Called by /ideate Phase 5 after /exp-pilot-run.

    Also in 4 projects that use this

  • skyllwtSkill

    exp-pilot-run

    Pilot experiment execution — read Pilot Spec YAML, write pilot code, run experiment(Confirm with the user before operation and require the applicant to conduct manual inspection), return results. Called by /ideate Phase 5. Does NOT modify wiki pages or judge pass/fail.

    Also in 4 projects that use this

  • +6 more from Lambenthan
  • +1 more from skyllwt

The file

Show its rules and text

We split this skill into the rules it gives an agent

19 rules. 0 are common, 4 are distinctive, and 15 are very specific to this project.

  • 0

    Common

    Most of the 74 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 15

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read the PDF (use Read tool.

    line 25
  2. Distinctive:

    Update biomarker files with timestamped rows in the body table:

    line 29
  3. Very specific:

    If any value triggers a diagnosis or category change (e.g., HbA1c crosses 5.7 → prediabetes, LDL crosses 100 → at-risk, B12 <200 pg/mL → …

    line 43
  4. Very specific:

    Update index.md with the new biomarker-event page.

    line 51
  5. Very specific:

    Update wiki/profile/medical.md.

    line 58
  6. Very specific:

    If new medication: ask user about timing/dose if not in report.

    line 59
  7. Very specific:

    If a follow-up appointment is mentioned, add to wiki/calendar/personal-events.md.

    line 60
  8. Very specific:

    Update wiki/biomarkers/sleep-recovery.md with new rolling averages.

    line 68
  9. Very specific:

    Update wiki/biomarkers/bp-resting-hr.md with RHR trend.

    line 69
  10. Very specific:

    Update wiki/profile/anthropometry.md if weight is synced and newer than the last manual entry (mark as "from Apple Health").

    line 70
  11. Very specific:

    Update wiki/biomarkers/sleep-recovery.md with daily rows + computed weekly averages (sleep avg, readiness avg, HRV trend).

    line 77
  12. Very specific:

    If readiness pattern is degrading (3+ low-readiness days in a week, or HRV trending down >10% over 2 weeks), surface in next checkin and …

    line 79
  13. Very specific:

    Update wiki/profile/inventory.md (add new items, increment existing items' last-purchased timestamps).

    line 89
  14. Very specific:

    Update wiki/biomarkers/body-composition.md with reference to the photo (relative link).

    line 97
  15. Distinctive:

    If user provided tape measurements alongside, log those too.

    line 98
  16. Distinctive:

    If a previous month's photo exists for the same view, describe visible changes qualitatively (do not over-claim.

    line 99
  17. Very specific:

    Look at content (Read for text/markdown/PDF, view for images).

    line 105
  18. Very specific:

    Use mv (don't copy; the file's destination is its real home).

    line 114
  19. Very specific:

    Update index.md if new pages were created.

    line 122

“Common” is measured: kept by at least 6 in 10 of the 74 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Ingest skill

Process any file in raw/, integrate findings into the wiki. NEVER modify the raw file.

File-type routing

Path / extensionSub-flow
raw/bloodwork/*.pdfBloodwork ingest
raw/medical/*.pdfMedical report ingest
raw/wearable-exports/*.zip (Apple Health)Apple Health ingest
raw/wearable-exports/*.csv (Oura)Oura ingest
raw/wearable-exports/*.csv (Apple Watch)Apple Watch ingest
raw/grocery-receipts/*.{jpg,pdf,png}Receipt OCR
raw/photos/body/*.{jpg,png}Body comp photo
raw/inbox/*Triage

Bloodwork ingest

  1. Read the PDF (use Read tool; pages parameter if PDF >10 pages).
  2. Identify lab provider (Thyrocare, Healthians, Lal PathLabs, Metropolis, etc.) from the header.
  3. Extract every value: name, unit, reference range, the user's value.
  4. Normalize units (e.g., glucose mg/dL vs mmol/L; B12 pg/mL vs pmol/L).
  5. Update biomarker files with timestamped rows in the body table:

    • Lipid panel → wiki/biomarkers/lipid.md
    • HbA1c, FPG, insulin, C-peptide → wiki/biomarkers/glycemic.md
    • LFT (ALT, AST, GGT, ALP, bilirubin) → wiki/biomarkers/liver.md
    • TSH, T3, T4, anti-TPO → wiki/biomarkers/thyroid.md
    • B12, Vit D, ferritin, iron, magnesium, folate → wiki/biomarkers/micronutrients.md
    • Creatinine, urea, eGFR, uric acid → wiki/biomarkers/kidney.md
    • CBC (Hb, WBC, platelets, RBC indices) → wiki/biomarkers/cbc.md

    In the SAME edit, update the YAML frontmatter at the top of that file (per Frontmatter contract in CLAUDE.md):

    • Append a new entry to the log: array with this reading's metrics.
    • Overwrite the latest: block with the most recent reading. The body and frontmatter must stay in lockstep. The Obsidian dashboard reads latest for the snapshot view.
  6. Compare against India thresholds in CLAUDE.md (LAI 2024 lipid, ICMR/ADA glycemic, India-specific micronutrient cutoffs). Flag abnormals.
  7. If any value triggers a diagnosis or category change (e.g., HbA1c crosses 5.7 → prediabetes, LDL crosses 100 → at-risk, B12 <200 pg/mL → deficient), update wiki/profile/medical.md and consider whether wiki/targets/current.md needs adjustment.
  8. Generate wiki/reviews/biomarker-events/YYYY-MM-DD-<labname>.md:
    • What changed since last reading (diff)
    • What's now flagged
    • What in the diet/lifestyle plan should adjust
    • Recommended physician follow-up if anything is genuinely abnormal
  9. Append to log.md: ## [DATE TIME] ingest | bloodwork from <lab>, <N abnormals>.
  10. Append to log.md: ## [DATE TIME] biomarker-event | <one-line summary>.
  11. Update index.md with the new biomarker-event page.
  12. Refresh hot.md with the key change.

Medical report ingest

  1. Read PDF.
  2. Extract diagnoses, recommendations, prescribed medications, follow-up dates.
  3. Update wiki/profile/medical.md. Cross-link to the source raw/ file (relative path).
  4. If new medication: ask user about timing/dose if not in report; update wiki/profile/supplements.md (or split out a wiki/profile/medications.md if the list grows beyond ~5 items).
  5. If a follow-up appointment is mentioned, add to wiki/calendar/personal-events.md.
  6. Append ## [DATE TIME] ingest | medical report from <doc/specialty> to log.md.

Apple Health ingest (zip)

  1. Acknowledge: "Apple Health export is large; I'll extract key signals."
  2. Parse the XML inside the zip (or ask user to extract export.xml for you to read).
  3. Pull: steps/day, resting heart rate, HRV, sleep stages and durations, weight (if synced), workouts.
  4. Update wiki/biomarkers/sleep-recovery.md with new rolling averages.
  5. Update wiki/biomarkers/bp-resting-hr.md with RHR trend.
  6. Update wiki/profile/anthropometry.md if weight is synced and newer than the last manual entry (mark as "from Apple Health").
  7. Append ## [DATE TIME] ingest | apple-health export, <date range> to log.md.
  8. Append ## [DATE TIME] sleep-ingest | <N nights> to log.md.

Oura ingest (csv)

  1. Parse Oura csv: sleep score, readiness score, HRV, RHR, body temp deviation, sleep stages, activity score.
  2. Update wiki/biomarkers/sleep-recovery.md with daily rows + computed weekly averages (sleep avg, readiness avg, HRV trend).
  3. Append ## [DATE TIME] ingest | oura, <date range> to log.md.
  4. If readiness pattern is degrading (3+ low-readiness days in a week, or HRV trending down >10% over 2 weeks), surface in next checkin and flag in hot.md.

Apple Watch ingest

Similar to Apple Health (same data sources usually). If Apple Watch data is in a separate file, merge with Apple Health on date keys.

Receipt OCR

  1. OCR the image/PDF.
  2. Extract item list with quantities and prices. Note store (Zepto, BigBasket, Blinkit, Reliance, etc.).
  3. Update wiki/profile/inventory.md (add new items, increment existing items' last-purchased timestamps).
  4. Flag items running low (compare to typical pantry levels in wiki/profile/kitchen.md).
  5. Compute meal cost contribution if helpful.
  6. Append ## [DATE TIME] ingest | receipt from <store>, <N items>, ₹<total> to log.md.

Body comp photo

  1. File the photo as raw/photos/body/YYYY-MM-{front,side,back}.jpg (rename if user dropped a generic filename).
  2. Update wiki/biomarkers/body-composition.md with reference to the photo (relative link).
  3. If user provided tape measurements alongside, log those too.
  4. If a previous month's photo exists for the same view, describe visible changes qualitatively (do not over-claim; vision can't measure body fat from a photo).
  5. Append ## [DATE TIME] ingest | body-comp photo, <view> to log.md.

Inbox triage

For each file in raw/inbox/:

  1. Look at content (Read for text/markdown/PDF, view for images).
  2. Decide destination:
    • Bloodwork PDF → move to raw/bloodwork/, then bloodwork ingest
    • Medical doc → move to raw/medical/, then medical ingest
    • Food photo → move to raw/photos/, then invoke log skill (meal-photo)
    • Body photo → move to raw/photos/body/, then body comp ingest
    • Receipt → move to raw/grocery-receipts/, then receipt OCR
    • Wearable export → move to raw/wearable-exports/, then matching ingest
    • Random text note → ad-hoc routing (could be a preferences update via checkin ad-hoc, a goal note, a symptom log via log, etc.)
  3. Use mv (don't copy; the file's destination is its real home).
  4. Invoke the appropriate sub-flow.
  5. Append ## [DATE TIME] ingest | inbox triage: <filename> → <destination> to log.md.

Bookkeeping (always)

Every ingest:

  1. Append to log.md with right kind.
  2. Update index.md if new pages were created.
  3. Refresh hot.md with the key change.
  4. Cross-link the source raw/ file in the wiki page (relative path) so the user can audit.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.