Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Request your medical records and actually get them — what to ask for, the request letter that can't be shuffled aside, the timelines and fee rules to cite (jurisdiction-flagged), and the escalation path for stonewalls. Use when asked how do I get my medical records, write a records request, my doctor's office won't send my records, or what records should I collect. Produces the itemized request letter, the delivery and format choices decoded, the follow-up ladder, and the personal health-file st
.claude/skills/mohitagw15856-medical-records-request/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 47% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 58% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 88% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 108% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 103% | 0% |
Your medical records are, in most places, yours by right — and yet getting them can feel like a heist: unanswered faxes, "we only send to other doctors," fees that materialize, imaging reports sent without the images. The gap is almost never law; it's process friction that a specific, itemized, deadline-aware written request cuts straight through. This skill writes that request, decodes the format choices that matter (images vs. reports, portal vs. complete file), and ladders the follow-up for offices that need reminding.
Ask for these if not provided:
Ready to send: patient identifiers · itemized records with date ranges · format and delivery choices · the generic access-rights sentence (verify-locally flagged) · fee-schedule request · the deadline if real · date and signature line]
| Item | Why this format | Common pitfall | |---|---|---|
Day 0: send via official intake, keep proof · Day ~10: polite status call, note the name · Day ~20: written reminder citing elapsed time and the request date · Beyond the local window: the complaint paths (records custodian's supervisor, the practice manager, and the applicable regulator — named as types, jurisdiction-flagged)
Structure for what arrives: by provider then date · the running summary page · what to hand the next new doctor — so this request is the last big one]
> Access rights, timelines, and fee caps are jurisdiction-specific and change — verify the local specifics before citing exact numbers; this skill's letters cite rights generically for exactly that reason.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 18,012 | 20,385 | +13% | 1 | 1 | 0% | 3,177 | 4,660 | +47% | 0 | 0 | — |
case-02 | fail→pass | 18,657 | 21,466 | +15% | 1 | 1 | 0% | 3,078 | 4,872 | +58% | 0 | 0 | — |
case-03 | fail→fail | 23,800 | 21,004 | -12% | 1 | 1 | 0% | 4,210 | 4,607 | +9% | 0 | 0 | — |
case-04 | pass→pass | 10,750 | 11,545 | +7% | 1 | 1 | 0% | 1,840 | 3,056 | +66% | 0 | 0 | — |
case-05 | pass→fail | 14,171 | 14,997 | +6% | 1 | 1 | 0% | 2,430 | 3,764 | +55% | 0 | 0 | — |
case-06 | pass→fail | 17,212 | 9,491 | -45% | 1 | 1 | 0% | 2,913 | 2,624 | -10% | 0 | 0 | — |
case-07 | pass→pass | 13,534 | 18,232 | +35% | 1 | 1 | 0% | 2,148 | 4,173 | +94% | 0 | 0 | — |
case-08 | fail→pass | 13,157 | 17,871 | +36% | 1 | 1 | 0% | 2,165 | 4,070 | +88% | 0 | 0 | — |
case-09 | fail→pass | 10,732 | 16,639 | +55% | 1 | 1 | 0% | 1,796 | 3,734 | +108% | 0 | 0 | — |
case-10 | fail→pass | 18,025 | 22,010 | +22% | 1 | 1 | 0% | 1,776 | 3,597 | +103% | 0 | 0 | — |
case-11 | fail→pass | 11,481 | 20,407 | +78% | 1 | 1 | 0% | 1,912 | 3,877 | +103% | 0 | 0 | — |
case-12 | fail→pass | 12,565 | 13,317 | +6% | 1 | 1 | 0% | 2,059 | 3,338 | +62% | 0 | 0 | — |
case-13 | pass→pass | 28,128 | 19,283 | -31% | 1 | 1 | 0% | 1,893 | 3,995 | +111% | 0 | 0 | — |
case-14 | fail→pass | 25,481 | 20,708 | -19% | 1 | 1 | 0% | 1,267 | 3,804 | +200% | 0 | 0 | — |
case-15 | pass→pass | 13,746 | 17,528 | +28% | 1 | 1 | 0% | 2,476 | 4,240 | +71% | 0 | 0 | — |
case-16 | pass→pass | 8,570 | 16,610 | +94% | 1 | 1 | 0% | 1,503 | 4,072 | +171% | 0 | 0 | — |
case-17 | fail→fail | 12,159 | 15,924 | +31% | 1 | 1 | 0% | 2,014 | 3,898 | +94% | 0 | 0 | — |
case-18 | fail→pass | 9,695 | 15,537 | +60% | 1 | 1 | 0% | 1,783 | 3,852 | +116% | 0 | 0 | — |
case-19 | fail→pass | 12,550 | 13,687 | +9% | 1 | 1 | 0% | 2,416 | 3,546 | +47% | 0 | 0 | — |
case-20 | pass→pass | 8,344 | 10,971 | +31% | 1 | 1 | 0% | 1,533 | 2,984 | +95% | 0 | 0 | — |
case-21 | pass→fail | 8,168 | 13,259 | +62% | 1 | 1 | 0% | 1,367 | 3,299 | +141% | 0 | 0 | — |
case-22 | pass→pass | 11,559 | 12,112 | +5% | 1 | 1 | 0% | 1,980 | 3,308 | +67% | 0 | 0 | — |
case-23 | fail→pass | 11,450 | 17,817 | +56% | 1 | 1 | 0% | 2,038 | 4,146 | +103% | 0 | 0 | — |
case-24 | fail→fail | 15,416 | 12,394 | -20% | 1 | 1 | 0% | 2,395 | 3,076 | +28% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 24 cases were attempted. The headline lift of +33 percentage points is the difference between those two pass rates over the 24 comparable cases. 3 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.