If you're trying to get your insurer to cover gender-affirming care in the United States, this project is for you. It's a set of Claude Code skills (with supporting scripts) that help you build the strongest possible case for coverage, file the right paperwork in the right places, and keep track of what you've sent and what comes next.
It can't promise an outcome. But it can help you to better advocate for yourself in a system that is designed to be almost impossible to navigate.
Mission: Find every means within a patient's power to get treatment, even when the system works against them.
Named after ICD-10 code F64 (gender dysphoria / gender identity disorders). Built by Juno Woods.
The skills are designed to be invoked from inside Claude Code. You describe your situation; the skill walks the agent through what to do.
| Skill | What it does for you | Status |
|---|---|---|
/f64:prior-auth |
Get a prior authorization filed correctly the first time. Figures out your plan type (PPO/HMO, in/out-of-network, ERISA self-funded), routes to the right intake at Anthem or Blue Shield, and sets follow-up reminders so nothing falls through the cracks. | Implemented |
/f64:gap-exception |
Get an out-of-network provider covered at in-network rates when no one in-network can treat you. Builds the three-prong argument (specialization, time, distance) and falls back to patient-direct submission if your provider won't help. | Implemented |
/f64:letter-check |
Check a mental-health practitioner's referral letter against WPATH SOC8 §12.1 before you submit it. Tells you what's missing and suggests language to add. | Implemented |
/f64:coverage |
Read your Evidence of Coverage adversarially — find what's actually covered, under what diagnosis code, with what documentation. Handles the silence-vs-exclusion games that some payers play. | Implemented |
/f64:appeals |
Navigate appeals, grievances, and DMHC complaints when a denial needs to be challenged. Walks the patient-direct path: written grievance, statutory window, regulator escalation, IMR. | Implemented |
Each implemented skill ships with evals.json (test cases the skill is graded against) and fixtures from real Evidence of Coverage documents.
The skills call into a small set of Python helpers — you generally don't run these directly, but they're here if you want to look:
| Script | Used by skills to... |
|---|---|
generate_prior_auth_fax.py |
Assemble a complete PA fax PDF — cover sheet, patient info, OON-exception language, attachments merged in. |
case_manager.py |
Track submitted PAs, follow-up dates, and communications locally. |
sinch_fax.py |
Send faxes via the Sinch Fax API v3 (no fax machine required). Requires a Sinch account with an activated project — copy .sinch.example to .sinch and fill in SINCH_KEY_ID, SINCH_KEY_SECRET, SINCH_KEY_PROJECT_ID. |
calendar_links.py |
Generate Google Calendar reminders for the deadlines that matter (5-business-day decisions, 30-day grievance escalation). |
/plugin marketplace add translunar/f64
/plugin install f64@f64
The f64 marketplace also hosts airbender (skill-authoring tooling) that you'll need if you want to assist with development:
/plugin install airbender@f64
git clone https://github.com/translunar/f64
cd f64
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt
pytest # runs tests for the helper scriptsThe skills are tuned to California's regulatory framework — DMHC, the Insurance Gender Nondiscrimination Act, the 5-business-day prior-auth decision deadline, the 30-day grievance/IMR escalation path. California has unusually strong patient protections, and the skills lean on them aggressively.
The project is actively expanding beyond California. If you're a patient or advocate working with a payer in another state, you're welcome here — see Contributing below. The full statutory scaffolding the California skills cite (and the common hallucinations they correct for) is in docs/california-context.md.
- Issue tracker: github.com/translunar/f64/issues — bugs, questions, gaps in payer coverage, "this didn't work for my plan," anything.
- Discord: discord.gg/tKuMsprBz7 — for in-the-moment questions and community support.
If you're navigating a denial right now and feeling stuck, opening an issue is a fine place to start. There's no wrong question.
Contributions are very welcome — especially from patients and advocates who've been through this themselves. The most useful contributions are usually:
- Firsthand experience with a payer the skills don't handle well. Kaiser, UnitedHealthcare, Aetna, Cigna, and smaller carriers all need more attention than they currently get.
- Out-of-state expansion. If you know your state's regulatory landscape, help us widen the project's scope.
- Real (de-identified) Evidence of Coverage documents to add to the fixtures.
- Letter templates, denial letters, and successful appeals that worked for you.
Open an issue or a PR. If you're not sure where to start, ask in an issue first — we'll help you find a useful first step.
- Not legal advice. The skills cite statutes and explain how to use them, but they don't replace a healthcare attorney for adversarial cases.
- Not a guarantee. Coverage is decided by the payer's clinical reviewer and (on appeal) by DMHC's Independent Medical Review process. The skills help you build the strongest case; outcomes depend on the case.
- Not exhaustive across payers. Anthem Blue Cross and Blue Shield of California are the most thoroughly handled. Kaiser, UnitedHealthcare, Aetna, and Cigna get reasonable coverage. Smaller carriers and out-of-state plans are weaker.
Experimental. Implemented skills are validated against synthetic and real EOC fixtures (Anthem Premier PPO 2024, Blue Shield Platinum Full PPO 2021, Kaiser KPIC PPO 2025) — see each skill's evals.json and fixtures/ directory.
MIT for all original work. See LICENSE for details.