pick-a-pii-model SKILL.md skillA

pick-a-pii-model from maziyarpanahi/openmed is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.

Install it, pinned
mdr add maziyarpanahi/openmed/pick-a-pii-model@git:20260727.9d7b5a2

Pins git:20260727.9d7b5a2 in mdr.lock. Pin the fingerprint instead.

Audit grade
A, 17 of 17 checks passed
Kind
SKILL.md skill
Repository
maziyarpanahi/openmed
Path
skills/pick-a-pii-model/SKILL.md
Versions recorded
1
Latest version
git:20260727.9d7b5a2, committed Jul 27, 2026
Size
3,192 bytes
Repository stars
5,784
License
Apache-2.0

Select an on-device OpenMed PII model from the committed registry by language, runtime format, and size budget, then require recall validation before deployment. Use when an agent must choose a local PII detector for CPU, Apple Silicon, or a mobile export without relying on live model discovery.

Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.

What the file says

# Pick an on-device PII model

Use the committed registry to build an offline shortlist. Treat the language
default as the safety baseline, but never treat model size or format as proof
of recall.

## Procedure

1. Identify the input language and script before choosing a model.
2. Choose the runtime: `pytorch` for local CPU/GPU and mobile export sources,
   `mlx-fp` or `mlx-8bit` for Apple Silicon.
3. Read `get_default_pii_model(language)` as the baseline.
4. Filter `get_pii_models_by_language(language)` by runtime and device budget.
5. Prefer the baseline when it fits; otherwise select a compatible candidate.
6. Benchmark the candidate on direct identifiers, critical leakage, scripts,
   and the target quantization before shipping.

## Runnable offline shortlist

This snippet reads only the bundled manifest; it does not download weights.

```python
from openmed import get_default_pii_model, get_pii_models_by_language

LANGUAGE = "en"
TARGET_FORMAT = "mlx-fp"  # Use "pytorch" for CPU or as an export source.
MAX_PARAMETERS_M = 150

baseline_id = get_default_pii_model(LANGUAGE)
models = get_pii_models_by_language(LANGUAGE)

shortlist = [
    (key, info)
…

Read the whole file at its exact version.

How to install

The exact contents, by fingerprint
mdr add maziyarpanahi/openmed/pick-a-pii-model@sha256:4e75f4547601339a

The line at the top of this page pins the label git:20260727.9d7b5a2, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.

Badge

mdr badge

[![mdr](https://markdownregistry.com/badge/art_zefatd4rka4zii5j.svg)](https://markdownregistry.com/a/art_zefatd4rka4zii5j)

3 badge image fetches in 30 days. What the badge shows.

Versions

The registry has recorded 1 version of this file, committed on Jul 27, 2026.

versioncommittedcommitsizeaudit
git:20260727.9d7b5a2 latest2026-07-27 9d7b5a2 3,192 BA view

Audit of the latest version

A  17 of 17 checks passed. Fixed automatic checks with no AI involved, so the same checks give the same answer every run.
  • pass: Frontmatter block present
  • pass: Frontmatter declares a name
  • pass: Frontmatter declares a description
  • pass: Size between 200 bytes and 200 KB (3192 bytes)
  • pass: No zero-width or bidi control characters
  • pass: No instruction hidden inside an HTML comment
  • pass: No link to an exfiltration or paste host
  • pass: No credential-shaped string
  • pass: No instruction to send local credentials anywhere
  • pass: No text hidden with inline styles
  • pass: No prompt-injection phrasing
  • pass: No curl or wget piped into a shell
  • pass: No recursive delete of root, home or parent
  • pass: No instruction to read or print local credentials
  • pass: No base64 blob over 200 characters
  • pass: No link to a raw IP address
  • pass: No script tag

Source

GitHub

maziyarpanahi/openmed · 5,784 stars · license Apache-2.0 · pushed 2026-11-07 · branch master

API

GET https://markdownregistry.com/api/v1/artifacts/art_zefatd4rka4zii5j
GET https://markdownregistry.com/api/v1/resolve?ref=maziyarpanahi/openmed/pick-a-pii-model
GET https://markdownregistry.com/api/v1/blob/4e75f4547601339a65455c683e67f485cd1f63e384d23d4975430499f0938cb8

Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.

Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.

More from maziyarpanahi/openmed

AGENTS.md AGENTS.md file
maziyarpanahi/openmed · AGENTS.md
git:20261102.e988f86 · audit A · 5,784 stars
annotating-variants SKILL.md skill
maziyarpanahi/openmed · skills/annotating-variants/SKILL.md · Annotates VCF variants and normalizes HGVS nomenclature with public, license-free annotators (Ensembl VEP REST…
v1.0 · audit A · 5,784 stars
ask-openmed SKILL.md skill
maziyarpanahi/openmed · skills/ask-openmed/SKILL.md · Choose the first OpenMed workflow skill for an intake, privacy, extraction, exchange, or verification request using…
v1.0 · audit A · 5,784 stars
assembling-fhir-bundles SKILL.md skill
maziyarpanahi/openmed · skills/assembling-fhir-bundles/SKILL.md · Package multiple FHIR R4 resources produced from OpenMed output into a single valid transaction Bundle ready to POST to…
v1.0 · audit A · 5,784 stars
auditing-deid-leakage SKILL.md skill
maziyarpanahi/openmed · skills/auditing-deid-leakage/SKILL.md · Adversarially scan already-de-identified clinical text for residual identifiers and emit a leakage report that blocks…
v1.0 · audit A · 5,784 stars
auditing-deidentification-runs SKILL.md skill
maziyarpanahi/openmed · skills/auditing-deidentification-runs/SKILL.md · Produce a signed, reproducible, no-PHI audit trail for an OpenMed de-identification run via deidentify(audit=True). Use…
v1.0 · audit A · 5,784 stars
auditing-part11-trails SKILL.md skill
maziyarpanahi/openmed · skills/auditing-part11-trails/SKILL.md · Generates and verifies 21 CFR Part 11-style audit trails — who/what/when, electronic signatures, and tamper-evidence —…
v1.0 · audit A · 5,784 stars
auditing-safe-harbor-checklist SKILL.md skill
maziyarpanahi/openmed · skills/auditing-safe-harbor-checklist/SKILL.md · Verify OpenMed de-identified output against all 18 HIPAA Safe Harbor identifier categories and report residual…
v1.0 · audit A · 5,784 stars
auditing-subgroup-fairness SKILL.md skill
maziyarpanahi/openmed · skills/auditing-subgroup-fairness/SKILL.md · Audit an OpenMed NER or de-identification model for performance disparities across demographic subgroups (sex, age…
v1.0 · audit A · 5,784 stars
authoring-model-cards SKILL.md skill
maziyarpanahi/openmed · skills/authoring-model-cards/SKILL.md · Generate a model card for an OpenMed clinical NER or de-identification model documenting intended use, quantitative…
v1.0 · audit A · 5,784 stars
batch-processing-clinical-text SKILL.md skill
maziyarpanahi/openmed · skills/batch-processing-clinical-text/SKILL.md · Run large-scale batch NER, PII extraction, or de-identification over many clinical notes on-device with OpenMed, with…
v1.0 · audit A · 5,784 stars
benchmark-pii-recall SKILL.md skill
maziyarpanahi/openmed · skills/benchmark-pii-recall/SKILL.md · Benchmark an OpenMed PII model with synthetic gold spans and report label-aware exact-span and grapheme recall without…
git:20260727.9d7b5a2 · audit A · 5,784 stars

Every file in maziyarpanahi/openmed

Browse by kind, by grade A, or by owner.