agentleFS
Open-MedicaOpen-Medica/open-medical-skills/skills/hipaa-compliance-checker/SKILL.md

hipaa-compliance-checker

Skill27 starsChanged 8 months agoMITView on GitHub

Clinical de-identification pipeline that detects 30+ PHI entity types. Provides masked and obfuscated output modes for HIPAA compliance. Runs locally with data never leaving secure environment.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Incident response: After an accidental data exposure, the compliance team runs the affected documents through the skill to catalog exactly …

    line 31

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “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

HIPAA Compliance & De-identification

A clinical de-identification pipeline capable of detecting and removing 30+ Protected Health Information (PHI) entity types from clinical text. The skill provides both masked output (replacing PHI with category tags like [PATIENT_NAME]) and obfuscated output (replacing PHI with realistic synthetic values) for HIPAA Safe Harbor and Expert Determination compliance. All processing runs locally with data never leaving the secure environment.

Quick Install

npx skills add Open-Medica/open-medical-skills --skill hipaa-compliance-checker

What It Does

  • 30+ PHI entity detection: Identifies all 18 HIPAA Safe Harbor identifiers plus additional clinical entities including patient names, dates of birth, medical record numbers, SSNs, phone numbers, email addresses, IP addresses, device identifiers, biometric identifiers, geographic subdivisions smaller than a state, and account numbers
  • Dual output modes: Provides masked mode (PHI replaced with bracketed category labels for audit trails) and obfuscated mode (PHI replaced with realistic synthetic data for research datasets that need to retain statistical properties)
  • Context-aware detection: Uses clinical NLP to distinguish between PHI and non-PHI uses of the same entity type (e.g., differentiating "Dr. Smith ordered the test" as a provider name from "the patient's brother Smith" as a relative identifier requiring de-identification)
  • Batch processing: Processes individual clinical notes or entire document collections, producing de-identified output alongside a PHI detection report with entity counts and confidence scores
  • Compliance reporting: Generates audit-ready reports documenting what PHI was detected, what action was taken, and the confidence level for each detection, supporting compliance officer review workflows

Clinical Use Cases

  • Research data preparation: A clinical researcher needs to share patient records with an external collaborator. The skill de-identifies an entire cohort's clinical notes, replacing all PHI with synthetic equivalents while preserving clinical content, temporal relationships, and data structure needed for the study
  • Training dataset creation: A health system building an internal NLP model needs de-identified clinical text for model training. The obfuscated output mode produces realistic-looking notes that preserve linguistic patterns without exposing real patient information
  • Incident response: After an accidental data exposure, the compliance team runs the affected documents through the skill to catalog exactly which PHI elements were present and generate a breach risk assessment
  • Cross-institutional data sharing: Two hospitals participating in a quality improvement collaborative use the skill to de-identify outcomes data before sharing, satisfying both HIPAA Safe Harbor requirements and their respective IRB protocols

Safety & Evidence

  • Safety Classification: Safe — The skill removes or replaces PHI; it does not transmit, store, or expose patient data. Processing occurs entirely within the local environment. The skill is a privacy-enhancing tool that reduces PHI exposure risk.
  • Evidence Level: High — De-identification methodology follows HIPAA Safe Harbor standards (45 CFR 164.514(b)(2)) with all 18 identifier categories covered. Clinical NLP entity detection is validated against i2b2/n2c2 de-identification benchmarks. The underlying engine has been tested across millions of clinical documents.

Example Usage

Masked output mode:

De-identify the following clinical note (masked mode):

"John Smith, DOB 03/15/1958, MRN 4472891, was seen in clinic today at
Massachusetts General Hospital. He reports his chest pain has improved
since his visit on 02/28/2026. Contact: (617) 555-0142."

De-identified output: "[PATIENT], DOB [DATE], MRN [MRN], was seen in clinic today at [FACILITY]. He reports his chest pain has improved since his visit on [DATE]. Contact: [PHONE]."

PHI Detection Report:

EntityTypeOriginalConfidence
John SmithPATIENT_NAMEJohn Smith0.99
03/15/1958DATE_OF_BIRTH03/15/19580.98
4472891MEDICAL_RECORD_NUMBER44728910.97
Massachusetts General HospitalFACILITYMGH0.96
02/28/2026DATE02/28/20260.99
(617) 555-0142PHONE_NUMBER(617) 555-01420.99

Obfuscated output mode:

De-identify the same note (obfuscated mode)

Obfuscated output: "Maria Chen, DOB 07/22/1961, MRN 7813254, was seen in clinic today at Riverside Community Medical Center. She reports her chest pain has improved since her visit on 01/15/2026. Contact: (312) 555-0897."

Technical Details

  • Category: Administrative
  • Author: OMS Contributors
  • License: Commercial
  • Version: 6.3.0
  • Specialty: Health Information Management, Compliance
  • Review Status: Verified by OMS Review Team

References

  • HIPAA Privacy Rule, 45 CFR 164.514(b) — Safe Harbor De-identification Standard
  • HHS Guidance Regarding Methods for De-identification of Protected Health Information (2012, updated 2022)
  • Stubbs A, Uzuner O, et al. "Automated systems for the de-identification of longitudinal clinical narratives: Overview of 2014 i2b2/UTHealth shared task Track 1." J Biomed Inform. 2015;58:S11-S19.
  • Dernoncourt F, et al. "De-identification of patient notes with recurrent neural networks." J Am Med Inform Assoc. 2017;24(3):596-606.
  • OCR HIPAA Breach Notification Rule, 45 CFR 164.400-414

This skill is part of Open Medical Skills, a curated marketplace of medical AI skills maintained by physicians for physicians and the healthcare industry.

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.