Security audit

Medical Research Toolkit

Security checks across malware telemetry and agentic risk

Overview

This is a documentation-only biomedical research skill that uses a disclosed external API endpoint, with privacy and medical-judgment cautions users should understand before using it.

Before installing, treat this as a research aid only. Do not submit patient identifiers, PHI, confidential study details, or proprietary hypotheses to the hosted endpoint unless you trust the operator and have appropriate data-handling approval. Protect OMIM or other API keys, avoid putting real keys in shared prompts or logs, verify the optional pip package before running it locally, and have qualified medical professionals validate any drug, diagnosis, safety, or treatment-related conclusions.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (6)

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: The README exposes a production biomedical MCP endpoint and provides copy-pasteable requests without warning users that submitted queries may contain sensitive research topics, proprietary target hypotheses, or patient-related information. In a medical research context, this can lead to unintended disclosure of confidential or regulated data to a third-party service, especially if users adapt examples with real clinical or investigational data.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The skill provides actionable drug discovery, safety, and clinical trial guidance without any disclaimer that outputs are informational only and not medical advice. In a medical-research context, users may over-trust returned drug and safety data and make clinical or self-treatment decisions without appropriate professional review, increasing risk of harm.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The checklist explicitly states that the production biomedical endpoint is public and unauthenticated, then provides copy-paste test commands without any warning against sending sensitive or regulated health-related data. In a medical research context, users may test with real patient cases, rare disease queries, or proprietary research terms, which could expose sensitive data to an external service without informed consent or handling guidance.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The file provides detailed genetic-disease analysis workflows and rare disease diagnosis guidance without a clear warning that OMIM content is reference material only and must not be used as a substitute for clinical diagnosis, genetic counseling, or patient-specific treatment decisions. In a medical research skill, this omission can encourage overreliance on database outputs for real patient interpretation, increasing the risk of misdiagnosis or inappropriate medical action.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: This documentation presents FDA adverse event and label data as part of a drug safety assessment workflow without clearly warning that FAERS reports are incomplete, non-causal, and not a substitute for clinician, pharmacist, or regulatory interpretation. In a medical research skill, users may over-trust the output for treatment or risk decisions, which can lead to unsafe clinical conclusions despite the included notes about under-reporting and label limitations.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: This workflow guides users through identifying non-standard-care drugs for a specific disease and ends with a recommendation template, but it does not clearly state that results are investigational research only and must not be used as clinical treatment advice. In a medical context, that omission can cause users to over-trust speculative repurposing outputs and make unsafe health decisions.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.