AI Claims Assessment
Check a healthcare-AI vendor's public claims against its own legal language and HIPAA
Public-document triangulation for healthcare-AI vendors. Reads a company's product pages, Terms of Service, and Privacy Policy, then surfaces where its market claims, legal commitments, and HIPAA's business-associate and de-identification rules fail to reconcile. Every gap is cited at both ends and framed as a public-documentation gap a covered entity would hit during vendor review.
From document to output
Paste or attach the vendor's product/marketing pages, Terms of Service, and Privacy Policy, or just one, if that's all you have
Every atomic claim is tagged by use (inference, model training, de-identification, and more) and data class (PHI, de-identified, or insufficient information)
Each use/data-class pair maps to its HIPAA requirement (business-associate permitted-use, de-identification method, or none) from a fixed lookup table
Market claims, legal language, and HIPAA requirements are checked against each other for divergence
Findings are grouped by Market↔Legal, Legal↔HIPAA, or Legal↔Legal, each cited at both ends, with a maturity tier
What you get back
Every finding is structured JSON: status, evidence, risk level, and remediation in one package. No interpretation required before acting on it.
{
"claim": "AI native system ... develop, improve, or demonstrate our Products and Services",
"source": "privacy",
"use": "Model training / model improvement",
"data_class": "insufficient public information to determine",
"gap_type": "legal-hipaa",
"selected_requirement": "Cross-document consistency + BA permitted-use (§164.502(a)(3), §164.504(e)(2)(i))",
"data_class_confidence": 0.9
} Three ways to run this skill
The first two are yours to run. The third is the platform, which runs as the instrument behind the practice rather than as a product sold here.
One-click install for Claude and Cowork users. The skill runs as a native tool in your Claude environment. No configuration, no MCP setup. Apache 2.0.
Get on ClawHub →Install as an MCP server in Claude Code, or download the SKILL.md and use it as a system prompt with any LLM. Runs the full methodology against documents you provide. Apache 2.0. No account required.
The platform runs this skill against a full document corpus in a workspace-isolated store: every finding is sourced to a specific document section, cross-referenced across the policy library, and carries a disposition and a reasoning trace. Not currently open for signups. It runs as the instrument behind the practice, and access is arranged inside an engagement.
How the practice runs it12 more skills in the methodology.
HIPAA Gap Analysis, BAA Review, Framework Mapping, Control Assessment, Risk Assessment, Part 2 Notice Assessment, Compliance Q&A, Compliance Posture Intake, Document Finder, Legal Page Discovery, Skill Router, and AI Tooling Inventory. All open source under Apache 2.0.
Need this done rather than explained? Dan Gonzalez runs Dang's Solutions, the practice that delivers it.