ClawHub
Back to skill

Security audit

Medication Safety Advisor

Security checks for vulnerabilities and agentic risk

Overview

This healthcare skill is mostly transparent, but it gives medication-change and dosing recommendations despite saying it is not for prescribing decisions.

Review before installing in clinical workflows. It should be limited to informational medication-safety support, with all recommendations checked by a licensed clinician or pharmacist, and users should avoid sending patient-identifying information or unapproved real patient data to public APIs.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

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 (3)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill explicitly says it should not be used for prescribing decisions, yet the example guidance gives concrete therapeutic actions such as switching antibiotics, holding statins, and making first-line treatment suggestions. In a healthcare context, this can cause users or downstream agents to rely on the skill for clinical decision-making beyond informational safety checks, creating patient safety and liability risk.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
90% confidence
Finding

The skill sends medication and allergy-related query data to an external public API, which is a real data egress path. Although the file warns not to include PHI, clinicians may still paste patient-specific medication lists or allergy details, so in practice this can expose sensitive health information to third parties if the warning is ignored or not enforced technically.

Content

Scanner excerpt · SKILL.md (reported line 201)May include surrounding context.

OpenFDA — Interactions & Labeling

text
GET https://api.fda.gov/drug/label.json?search=openfda.brand_name:{name}
→ Returns full drug labeling including interactions section

GET https://api.fda.gov/drug/event.json?search=patient.drug.medicinalproduct:{name}

External Transmission

Medium
Category
Data Exfiltration
Confidence
89% confidence
Finding

This second OpenFDA endpoint is another explicit external transmission channel and may send drug names or related patient context to a public service. In a healthcare workflow, repeated external calls increase the chance that sensitive medication or allergy information is disclosed outside approved compliance boundaries, especially if users submit raw patient notes.

Content

Scanner excerpt · SKILL.md (reported line 204)May include surrounding context.

GET https://api.fda.gov/drug/label.json?search=openfda.brand_name:{name} → Returns full drug labeling including interactions section

GET https://api.fda.gov/drug/event.json?search=patient.drug.medicinalproduct:{name} → Returns adverse event reports

text

Static analysis

No suspicious patterns detected.