Back to skill

Security audit

XML读取器免费版

Security checks for vulnerabilities and agentic risk

Overview

This XML reader skill is mostly a local Markdown instruction package, but it inconsistently claims to be read-only while requesting write access and describing create, modify, delete, save, and export behavior.

Install only if you are comfortable reviewing each command before it runs. Treat the free edition as ambiguously scoped: use it for local XML reading and querying, and avoid write, save, export, modify, delete, reset, or import operations unless the publisher clarifies the behavior and destination controls.

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
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • 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 (4)

Intent-Code Divergence

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill repeatedly states the free edition is read-only, but the capability sections advertise operation modes such as create, modify, delete, reset, import, save, and export. This mismatch can cause an agent or user to invoke file-changing behaviors under the false assumption that the tool is non-destructive, increasing the risk of unauthorized modification or deletion of local data.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill's display name, summary, and the entire user-facing documentation are presented in Chinese only, which imposes a specific language/locale on users. The file does not offer an alternative language, localization setting, or explicit opt-in for Chinese-only behavior.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The examples document output/save/export behavior to files, while the FAQ says the free version only outputs to the terminal. This inconsistency may mislead an agent into writing files unexpectedly, creating side effects on the local filesystem despite the user's expectation of a non-persistent read-only tool.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The markdown includes commands and options implying save/export/file-output behavior without a clear safety warning that local files may be created or overwritten. In an agent context with read, write, and exec tools available, ambiguous documentation about write-side effects increases the chance of unintended filesystem changes.

Content

No source excerpt is available for this finding.