Ae4
Medium
- Category
- analysis-evasion
- Confidence
- 80% confidence
- Finding
Suspicious Unicode normalization or mixed-script content
- Content
Security audit
Health Checkup Report
Security checks for vulnerabilities and agentic risk
Overview
The skill is not malicious, but it should be reviewed because its medical examples sometimes move from report explanation into diagnosis and treatment advice.
Review this skill carefully before installing if it will be used by non-clinicians. It appears safe from a code, persistence, and data-exfiltration perspective, but users should treat its outputs as educational report explanations only and seek licensed medical care for abnormal, symptomatic, urgent, or treatment-related decisions.
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
Vulnerability Patterns
- Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
- Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
- MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
- Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
- Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (5)
Intent-Code Divergence
Medium
- Category
- Not specified by scanner
- Confidence
- 96% confidence
- Finding
The skill states it is not for diagnosis, but later instructions and examples encourage the model to present disease conclusions and management recommendations as if they were clinical determinations. In a medical-report interpretation context, this contradiction can push the agent to provide unlicensed medical advice, creating safety risk through false reassurance, delayed care, or inappropriate self-treatment.
- Content
Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 95% confidence
- Finding
The file explicitly instructs the skill to 'Use plain Chinese,' which forces a specific language/locale for all outputs. This is a natural-language policy concern because the skill does not mention any user opt-in or alternative language handling.
- Content
Intent-Code Divergence
Medium
- Category
- Not specified by scanner
- Confidence
- 97% confidence
- Finding
This sample output explicitly concludes 'most likely iron deficiency anemia' and recommends confirmatory testing plus iron treatment, which goes beyond informational interpretation into diagnosis and treatment advice. Because examples strongly steer model behavior, this increases the chance the deployed agent will give actionable medical recommendations without adequate clinical context or safeguards.
- Content
Intent-Code Divergence
Medium
- Category
- Not specified by scanner
- Confidence
- 97% confidence
- Finding
The example presents hyperthyroidism as a concluded diagnosis and gives specialty workup guidance, despite the skill's earlier rule that it should not diagnose disease. In a user-facing health context, this can cause overconfidence in an AI-generated conclusion and may influence urgent decisions, medication expectations, or avoidance of appropriate in-person evaluation.
- Content
Static analysis
No suspicious patterns detected.