Back to skill

Security audit

XML转JSON工具

Security checks for vulnerabilities and agentic risk

Overview

The skill appears to be a broad, under-scoped automation template rather than a tightly limited XML-to-JSON converter.

Review before installing. This skill does not show evidence of a hidden payload, but it asks for broad agent powers and credentials while giving vague, inconsistent instructions. Only use it in a constrained environment if you specifically need its Expanso pipeline behavior, and avoid giving it sensitive XML, API keys, shell access, or broad file-write authority unless the publisher clarifies exact commands, network destinations, and data handling.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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
Findings (8)

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
95% confidence
Finding

Advertising command execution for a simple XML-to-JSON use case is unjustified and materially increases risk. If an agent is allowed to follow this skill, exec capability can be abused to run arbitrary system commands, leading to code execution, environment access, file modification, or lateral impact on the host.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The invocation guidance is overly broad and overlaps with generic coding, debugging, deployment, and development-help requests. That makes accidental triggering more likely in unrelated conversations, potentially causing the agent to choose an overprivileged skill when a simpler, safer tool would suffice.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill is presented as a narrow XML-to-JSON converter, but its declared tools and surrounding text imply broader automation behaviors including read/write/exec. This mismatch increases the chance that an agent or user invokes it with more privilege than necessary, enabling unintended filesystem or command-side effects beyond the stated purpose.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The inclusion of unrelated premium features such as visualization, scheduled sync, and data cleaning indicates the documentation may be templated or misleading rather than specific to this skill. While not an exploit by itself, misleading capability claims can hide actual behavior and undermine safe review, approval, and user consent.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The skill is framed as a focused converter, but later sections broaden it into file handling, API integration, command execution, and information retrieval. This inconsistency is dangerous because users and orchestrators may grant trust based on the narrow description while the effective behavior is much more powerful.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill mentions file writing, API calls, and command execution without prominent upfront warning that it can affect the system or transmit data. This weakens informed consent and increases the chance that an agent executes impactful actions under the guise of a harmless format-conversion utility.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

External API integration and general information retrieval are not necessary for local XML-to-JSON conversion and expand the trust boundary without clear justification. This can expose input data to third parties, enable data exfiltration, or cause the skill to fetch and act on untrusted remote content.

Content

No source excerpt is available for this finding.

Vague Triggers

Low
Category
Not specified by scanner
Confidence
89% confidence
Finding

The setup steps instruct the user to call the skill in an AI Agent conversation but do not define how the skill should be invoked or what requests should not trigger it. For a markdown skill description, this leaves the activation boundary unclear.

Content

No source excerpt is available for this finding.