Back to skill

Security audit

Med-LLM-Wiki

Security checks for vulnerabilities and agentic risk

Overview

The skill is purpose-aligned but needs review because it stores and indexes sensitive patient health details without privacy or retention safeguards.

Install only if you have authority to process the patient data and can keep the generated wiki in a protected location. De-identify records before ingestion, avoid public or shared repositories, exclude generated patient files from source control, and define deletion and retention practices before using it with real cases.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (1)

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.md:201
Finding

Persistent Storage and Broad Indexing of Sensitive Patient Health Data

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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)

Missing User Warnings

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill is explicitly designed to process real patient records, clinical consultations, and medical Q&A, but it provides no privacy warning, consent requirement, minimization guidance, or restriction on handling protected health information. In a medical context, silently encouraging ingestion of identifiable health data creates a substantial risk of collecting, retaining, and redisclosing highly sensitive information.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

These instructions direct the agent to create and update persistent patient-specific wiki pages containing demographics, diagnoses, test values, consultation questions, and treatment details. Persisting this material to structured files without safeguards, access controls, or warnings materially increases the chance of long-term exposure, secondary use, or unauthorized access to sensitive medical data.

Content

No source excerpt is available for this finding.

Ssd 3

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill instructs preservation of all numerical values, units, baseline measurements, and per-patient consultation content, then indexes that information for retrieval. That design increases re-identification and data leakage risk because highly specific medical facts, timelines, and linked records can expose an individual's health status even if some obvious identifiers are omitted.

Content

No source excerpt is available for this finding.