Discharge Summary Writer

Security checks across malware telemetry and agentic risk

Overview

This skill locally drafts discharge summaries from user-provided patient data and discloses the clinical-review and PHI-handling risks.

Install only in an environment approved for patient data. Treat generated summaries as drafts, verify all diagnoses, medications, dates, follow-up instructions, and language with a qualified clinician, and choose output paths carefully because the script can overwrite the selected file.

SkillSpector

By NVIDIA

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (2)

Vague Triggers

Medium

Confidence: 80% confidence
Finding: The manifest description uses broad trigger language such as compiling inpatient medical records and generating post-hospitalization documentation, which can cause the skill to activate for requests beyond narrow discharge summary generation. In a clinical context, overbroad routing can expose highly sensitive patient data to the wrong workflow, produce inappropriate documentation, or bypass more suitable, better-controlled tools.

Natural-Language Policy Violations

Medium

Confidence: 76% confidence
Finding: Defaulting output language to Chinese without explicit user or institutional selection can create unsafe clinical documentation by generating discharge instructions in a language the patient, clinician, or receiving system may not expect. In healthcare, language mismatch can directly affect medication understanding, follow-up compliance, and legal/documentation correctness.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal