ClawHub
Back to skill

Security audit

Tone Adjuster

Security checks for vulnerabilities and agentic risk

Overview

This skill is a simple medical text tone-adjustment helper with no hidden access, persistence, or data exfiltration, but its documentation overstates some APIs that are not present in the packaged script.

Before relying on this skill, treat it as a basic local text-rewriting helper rather than a clinically validated medical communication tool. Review outputs for medical accuracy, and use the packaged scripts/main.py interface rather than the undocumented examples that reference missing APIs or files.

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 (1)

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill documentation advertises Python APIs and capabilities such as ToneAdjuster, to_patient_friendly, to_academic, and reading-level/jargon functions, while the declared executable path elsewhere is scripts/main.py. If those APIs are not actually implemented, operators or downstream agents may rely on nonexistent functionality, causing unsafe automation assumptions, incorrect medical-content transformations, or execution of ad hoc fallback code outside the reviewed path. In a medical-text skill, misleading capability claims are more dangerous because users may trust the output for patient-facing communication.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.