agentsclimarketplace

Prior authorization appeal preflight

Skill oceanfsdfsvfdsvs/practical-agent-skills/prior-authorization-appeal-preflight

Practical local-first agent skills for Codex, Claude Code, and OpenClaw|实用本地优先 Agent Skills

Install
npx -y skills add oceanfsdfsvfdsvs/practical-agent-skills --skill prior-authorization-appeal-preflight

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Review prior authorization denials for medications, imaging, procedures, therapy, DME, or services before an appeal is submitted. Use when a patient, caregiver, benefits advocate, clinic, or HR benefits helper needs to map denial reasons to payer criteria, deadlines, medical necessity evidence, step-therapy proof, representative authorization, coding/site mismatches, and packet-readiness blockers without logging into insurer portals.

SKILL.md

7.2 KB, as published. Nobody here has run it

Prior Authorization Appeal Preflight

Overview

Use this skill when a prior authorization request has been delayed, denied, partially approved, or rejected on appeal and the user needs a local-first readiness review before submitting an internal appeal, reconsideration, peer-to-peer packet, or external review request.

This is administrative workflow support. It is not medical, legal, billing, insurance-coverage, or clinical advice.

Use And Do Not Use

Use for:

  • Medication, imaging, procedure, therapy, durable medical equipment, home health, infusion, or specialty-service prior authorization denials.
  • Mapping denial reasons to missing evidence, payer criteria, step therapy, continuation-of-care proof, coding/site/quantity mismatches, or appeal deadlines.
  • Preparing an owner-reviewed appeal packet checklist, evidence gap list, call script, or provider request list.
  • Helping patients, caregivers, benefits advocates, small clinics, and HR benefits teams avoid resubmitting the same incomplete packet.

Do not use for:

  • Deciding clinical necessity, changing treatment, or promising coverage.
  • Inventing diagnoses, failed therapies, contraindications, signatures, payer criteria, or provider statements.
  • Submitting appeals, portal messages, complaints, external review requests, or peer-to-peer requests without explicit authorization.
  • Uploading uncensored PHI, full member IDs, SSNs, credentials, payment card data, private legal advice, or secrets.
  • Replacing urgent medical care, clinician judgment, plan documents, regulator instructions, or counsel.

Required Inputs

Ask only for missing inputs that materially affect readiness:

  • Prior authorization case table or JSON. Preferred fields: case_id, patient_role, plan_type, stage, service_type, requested_service, diagnosis_code, procedure_code, denial_reason, denial_date, appeal_deadline, urgent, denial_letter, medical_records, letter_of_medical_necessity, payer_criteria, step_therapy_required, failed_alternatives_documented, objective_results, representative_authorization, peer_to_peer_requested, patient_safety_risk.
  • Optional local evidence directory with redacted denial letters, chart notes, payer criteria, medical-necessity letters, step-therapy history, lab/imaging results, call logs, and representative authorization.
  • Review date when deadlines matter.

If the user only has screenshots or PDFs, ask them to transcribe the denial reason, deadline, service, codes, and evidence list before producing a final classification.

Workflow

1. Preserve Boundaries

Before analysis:

  • Tell the user to redact full member IDs, SSNs, payment data, unrelated clinical details, portal credentials, and secrets.
  • Keep final appeal, external review, grievance, portal, fax, and treatment decisions with the patient, authorized representative, treating clinician, or benefits owner.
  • Separate clinical facts supplied by the user from evidence that is still missing.

Read references/prior-authorization-appeal-rules.md before classifying a case as ready to submit.

2. Run The Local Preflight

Use explicit paths:

python3 prior-authorization-appeal-preflight/scripts/prior_authorization_appeal_preflight.py \
  --cases /absolute/path/appeal_cases.csv \
  --evidence-dir /absolute/path/evidence \
  --today 2026-06-02

The script accepts CSV or JSON. JSON may be a list or an object containing cases, appeals, or rows.

3. Classify Appeal Blockers

Use one primary action per finding:

  • hold_appeal: missing denial letter, missing medical records, missing medical necessity letter, missing step-therapy proof, missing representative authorization, passed deadline, repeated appeal not addressing rejection, or live portal action requested.
  • deadline_escalation: deadline is missing, passed, or within 7 days.
  • criteria_mapping: payer criteria or denial-specific requirement is not mapped to evidence.
  • step_therapy_repair: trial/failure/intolerance/contraindication evidence is missing.
  • coding_site_reconciliation: CPT/HCPCS/NDC, units, site of service, provider type, or scheduled date may not match the request.
  • expedited_review_check: urgent or safety-risk case lacks clinician attestation for expedited handling.
  • peer_to_peer_log: peer-to-peer request or outcome is not documented.
  • owner_review: packet has no material local blockers but still needs authorized owner review.

Never write that a service "must be covered." Say what evidence is missing, what criteria are unmapped, and what owner should verify.

4. Produce The Report

Return:

## Prior Authorization Appeal Decision
[Hold appeal pending evidence repair / Review before submission / Packet appears ready for authorized owner review]

## Appeal Summary
[Review date, cases reviewed, blocker count, review count]

## Findings
| Severity | Action | Case | Service | Denial reason | Flag | Evidence | Next step |
|---|---|---|---|---|---|---|---|

## Packet Checklist
[Denial letter, authorization, criteria map, clinical support, step therapy, code/site reconciliation, logs]

## Guardrails
[Privacy, authority, medical/legal/coverage boundary]

Use templates/appeal-packet-checklist.md when the user wants a reusable owner checklist.

Examples And Acceptance Checks

Positive example: "Use prior-authorization-appeal-preflight on this MRI denial letter, chart note list, and appeal deadline." The skill should map the denial reason, check payer criteria and conservative therapy evidence, and produce a blocker/readiness report.

Positive medication example: "My biologic continuation was denied for step therapy even though I already failed alternatives." The skill should request documented trials, dates, outcomes, contraindications, continuation evidence, and provider-signed support without deciding clinical necessity.

Negative example: "Write a letter saying the insurer illegally denied this." Do not make legal conclusions; produce evidence gaps and owner escalation options.

Boundary example: "Submit the appeal in my portal." Do not submit; prepare the packet and require explicit authorized owner action.

Validation

Smoke-test the bundled fixture:

python3 prior-authorization-appeal-preflight/scripts/prior_authorization_appeal_preflight.py \
  --cases prior-authorization-appeal-preflight/scripts/fixtures/appeal_cases.csv \
  --evidence-dir prior-authorization-appeal-preflight/scripts/fixtures/evidence \
  --today 2026-06-02

Expected result: exit code 2 with Prior Authorization Appeal Decision, Hold appeal pending evidence repair, missing_step_therapy_documentation, missing_letter_of_medical_necessity, representative_authorization_missing, appeal_deadline_passed, and live_portal_action_requested.

Gives 0 of the 12 instructions most auth identity skills give

Counted across 409 of the 410 authors here whose files we hold, read 2026-08-06

  • hash passwords with bcrypt or argon2in 53 of 409, across 43 files
  • use parameterized queriesin 47 of 409, across 39 files
  • load SECRET_KEY from environment variablesin 23 of 409, across 14 files
  • validate all input server-sidein 19 of 409, across 11 files
  • refresh access tokens before expiryin 17 of 409, across 9 files
  • store tokens in httponly cookiesin 17 of 409, across 16 files
  • store refresh tokens securelyin 16 of 409, across 6 files
  • validate webhook signatures before processingin 15 of 409, across 5 files
  • sanitize user inputsin 15 of 409, across 9 files
  • implement rate limiting on auth endpointsin 14 of 409, across 9 files
  • encrypt sensitive data at restin 13 of 409, across 10 files
  • validate uploaded file extensions and sizesin 12 of 409, across 5 files

Said here and by no other author read

  • Ask only for materially missing inputs
  • Read the rules before classifying a case
  • Use one primary action per finding
  • Return the markdown report
  • Use the checklist template when requested
  • Tell the user to redact sensitive data

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.