Security audit

域名DNS管理免费版

Security checks across malware telemetry and agentic risk

Overview

This DNS management skill is mostly coherent, but it mixes high-impact DNS-changing workflows with unrelated project-management trigger text that could cause unsafe invocation.

Review this before installing because it can guide an agent to make real DNS and nameserver changes. Only use it for explicit domain/DNS tasks, remove or ignore the unrelated project-management trigger language, keep provider tokens least-privilege, and require manual confirmation plus a backup of current DNS records before any live change.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
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
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (6)

Description-Behavior Mismatch

Medium

Confidence: 97% confidence
Finding: The manifest description says this is a DNS/domain management skill, but also includes unrelated project-management usage guidance. This mismatch can cause the agent to invoke a high-impact infrastructure skill in irrelevant contexts, increasing the chance of accidental DNS or nameserver changes. Because the skill exposes exec/write-capable operational workflows, incorrect routing is materially risky rather than merely a documentation bug.

Intent-Code Divergence

Medium

Confidence: 95% confidence
Finding: The capability coverage section claims support for project-management scenarios and many unrelated keywords even though the documented functions are limited to DNS/domain operations. This can broaden matching and selection logic, causing an agent to choose a privileged infrastructure-modification skill for unrelated user requests. In practice, that raises the risk of unintended execution paths involving DNS changes, redirects, or registrar operations.

Vague Triggers

High

Confidence: 98% confidence
Finding: The description contains broad, mismatched trigger text ('project management', 'task planning', 'team collaboration') for a skill that can drive operational DNS changes. Overbroad invocation criteria are dangerous here because the skill has exec/write tooling and documents commands that affect live internet-facing infrastructure. A mistaken invocation could lead to outage-inducing nameserver changes or record updates under the wrong user intent.

Vague Triggers

Medium

Confidence: 95% confidence
Finding: The capability scope enumerates excessively broad keywords, including generic words like 'Use' and unrelated business/workflow phrases. In a skill-selection pipeline, such generic terms can act like catch-alls and cause this infrastructure-management skill to be invoked outside its proper context. Given the operational nature of the commands, even accidental selection can have user-visible service impact.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The documentation gives direct instructions for nameserver switching and DNS record modification but does not prominently warn that these are service-impacting operations that can break websites, email delivery, and verification flows. In a skill intended for automated agent use, omission of explicit pre-change confirmation and rollback safeguards materially increases the chance of outages. The dangerousness is amplified because nameserver changes can redirect full-zone resolution, not just a single record.

Missing User Warnings

Low

Confidence: 84% confidence
Finding: The skill explains how to configure provider API tokens but does not clearly warn about secret exposure risks, least-privilege scoping, logging leakage, or avoiding echoing credentials into shells/history. While it advises using environment variables and forbids hardcoding, the absence of stronger credential-handling guidance can still lead to accidental token disclosure in agent logs, transcripts, or shared environments. In this context, exposed DNS-provider credentials could allow unauthorized DNS changes.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.