Back to skill

Security audit

防火墙

Security checks for vulnerabilities and agentic risk

Overview

This firewall automation skill is not malicious, but it gives broad command/write authority and vague instructions for high-impact network changes.

Review before installing. Use this skill first in audit or dry-run workflows, require approval before any allow/deny change, keep console or rollback access available, and limit credentials, hosts, cloud projects, and writable files to the specific firewall task.

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
  • 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 (5)

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The skill says it is limited to network firewall use, but other sections advertise unrelated file-processing and automation capabilities. This inconsistency broadens the apparent authority of the skill and can cause an agent to use read/write/exec powers for tasks outside the declared scope, which raises the risk of over-privileged or unintended actions.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The documented JSON result format describes generic code-style and compliance scoring rather than firewall rules, security groups, ports, or audit findings. This mismatch can mislead an agent or user about what the skill actually does, increasing the chance of unsafe automation, incorrect downstream parsing, or applying the skill in the wrong operational context.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

A skill presented as a firewall tool but documented as a generic file-processing, API-integration, and command-execution utility creates scope ambiguity around powerful capabilities. In a security-sensitive context, that ambiguity is dangerous because agents may grant or exercise broad operational behavior under the trusted label of firewall management.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill advertises automated firewall configuration and command execution but does not prominently warn that live changes can lock out administrators, disrupt services, or expose systems if rules are applied incorrectly. Because firewall changes affect network reachability and security posture directly, insufficient safety gating materially increases operational risk.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
81% confidence
Finding

The skill presents its display name and substantial descriptive content in Chinese, and the document predominantly instructs users in Chinese without stating that another language can be used. Under the policy, locale or language constraints should be optional or explicitly justified rather than imposed by default.

Content

No source excerpt is available for this finding.