Back to skill

Security audit

English Word Coach

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed vocabulary coach that stores review progress in local Markdown files and does not show deceptive, destructive, or privilege-seeking behavior.

Install this if you want a Chinese-language English vocabulary coach that keeps progress files in the current project. Be aware that entering a single English word can add it to local memory files, so avoid using it from repositories or synced folders where you do not want vocabulary history written.

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
  • 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 (2)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The instructions require output in a fixed language format: 'Chinese explanations + English examples', and later sections continue using Chinese-only labels and reminders. This forces a specific language/locale behavior without presenting it as an option or obtaining user preference, which matches the language-policy violation criteria.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill is explicitly designed to create and update files under memory/ in the current working directory, but it provides no user-facing notice or consent step before persisting local data. This can lead to unexpected modification of the user's workspace, accidental storage of sensitive inputs, and confusion in environments where the working directory is shared, synced, or under version control.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.