DNA
WarnAudited by ClawScan on May 18, 2026.
Overview
This instruction-only skill asks the agent to use highly sensitive DNA and health data while making strong privacy and clinical-grade claims that are not backed by the supplied artifacts.
Treat this as a Review item before installing. Do not upload raw DNA, VCF/FASTQ files, medication lists, or other health data unless you have independent confirmation of where processing occurs, how data is protected and deleted, and whether a qualified clinician should review any recommendations.
Findings (2)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
A user may be led to trust privacy protections or clinical-grade genomic conclusions that the reviewed artifacts do not actually demonstrate.
The skill makes strong privacy and clinical-grade capability claims, but the supplied package is instruction-only with no code, install mechanism, enclave implementation, or database provenance to substantiate them.
"privacy": "Zero-Knowledge Processing — DNA never leaves the encrypted enclave" ... "logic": "Multi-variate SNP analysis across 2026 clinical grade databases"
Do not rely on the privacy, enclave, or clinical-grade claims unless the publisher provides auditable implementation details, data sources, retention controls, and medical-use limitations.
Raw DNA files and derived health markers could enter the agent's context and potentially be reused or exposed in ways the user did not intend.
The skill explicitly contemplates raw genomic data being interpreted by the agent and used to influence decisions, but it does not clearly bound data scope, retention, reuse, output handling, or user approval for such sensitive context.
"ingestion": "Direct interpretation of raw FASTQ/VCF genomic data" ... "The bridge between your biological identity and your agent's decision-making engine."
Avoid providing raw genomic files or prescription-related details until the skill clearly documents local-only processing, retention/deletion behavior, allowed outputs, and user-controlled consent boundaries.