ClawHub
Back to skill

Security audit

iaiops-clinical

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly transparent, but it describes a hospital facilities skill as read-only while also documenting write-capable building-control operations.

Review this before installing in any hospital or facility-control environment. The package is not malicious code, but the documentation mixes read-only framing with live write-capable BACnet behavior. Only use it where write tools are separately permissioned, approval-gated, and blocked from life-safety or patient-safety points unless your site has explicit MOC authorization.

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
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The manifest and description repeatedly frame this edition as read-only, but the documented tool surface includes bacnet_write_property, a live write capability to BACnet building controls. In a clinical environment, that mismatch is dangerous because users or higher-level orchestration may trust the skill as non-actuating and invoke it in sensitive contexts involving isolation-room pressure or medical-gas related systems.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The skill labels Modbus and OPC-UA sections as read-only, yet the supported-version matrix advertises Modbus function codes including 5/6/15/16, which are write operations. Even if the current exposed tools are documented as read-only, this inconsistency can mislead operators, auditors, or automated policy systems about the actual actuation potential of the stack.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.