Security audit

link-skills

Security checks across malware telemetry and agentic risk

Overview

This is a coherent Link project development skill, but it exposes sensitive internal credentials/configuration and gives live operations guidance without enough safety controls.

Review and sanitize this skill before installing: rotate the exposed Nexus credential and AES key, replace internal endpoints and secrets with placeholders, add explicit production approval and rollback gates for deployment/ops steps, and limit signature-bypass or delete examples to isolated local test environments.

SkillSpector

By NVIDIA

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (9)

Context-Inappropriate Capability

Medium

Confidence: 99% confidence
Finding: The document exposes a hardcoded AES key and concrete internal infrastructure details, including database and Kafka endpoints. Even though this is 'just documentation,' such secrets and topology data materially lower the bar for unauthorized access, lateral movement, and offline decryption of protected configuration values.

Intent-Code Divergence

Medium

Confidence: 86% confidence
Finding: The example is presented as safe, but it normalizes dynamic SQL construction patterns around user-controlled ordering and contains malformed MyBatis control flow that could be copied incorrectly into production code. In a development skill, unsafe or ambiguous secure-coding examples are dangerous because they are likely to be reused by downstream agents or developers without rigorous validation.

Context-Inappropriate Capability

Medium

Confidence: 97% confidence
Finding: The guide explicitly recommends disabling request-signature verification in development by setting `sign-check: false`, which weakens an authentication/integrity control without any guardrails or warning about the risk. Even if intended only for local use, such guidance is often copied into shared, test, or accidentally deployed environments, creating a path for unsigned requests to reach the service.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill includes detailed deployment, rollout, configuration-change, and operational troubleshooting instructions for live services, but it does not require explicit confirmation, environment verification, backups, rollback planning, or a warning that these actions may affect production systems. In an agent context, this omission can cause unsafe execution of Helm upgrades, config changes, or operational remediation against real infrastructure with service disruption, data integrity, or availability impact.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The guide documents a 'data relay' pattern where requests and data are sent to a third-party knowledge-base API, but it does not clearly warn operators or users that data may leave the local environment. In an enterprise development/deployment skill, this can lead to unintended external transmission of sensitive business data, compliance violations, and mistaken assumptions about data residency.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: This section instructs users to disable signature verification but does not clearly state that doing so removes a security control and must never be used beyond isolated local development. Documentation that normalizes insecure configuration increases the likelihood of unsafe reuse in broader environments.

Missing User Warnings

Medium

Confidence: 89% confidence
Finding: The deletion test example performs destructive operations (`deleteById`) without warning that it will remove real data from the target system. In a testing guide for operational services, this can lead to accidental data loss if run against shared or production-like environments.

Missing User Warnings

Medium

Confidence: 96% confidence
Finding: The guide explicitly instructs developers to disable inbound signature verification in application-dev.yml, which weakens an authentication/integrity control. Even if framed as local troubleshooting, this is dangerous because such settings are often copied into shared environments, committed accidentally, or used as a template for broader deployments, enabling unauthorized requests and bypassing request validation.

Missing User Warnings

High

Confidence: 99% confidence
Finding: The document contains a plaintext Nexus repository username and password in a sample settings.xml snippet, which is a direct credential exposure. In the context of a development/operations skill for a real microservice environment, this is especially dangerous because readers may reuse the credential immediately, and anyone with access to the skill content may gain unauthorized access to internal artifact repositories or pivot further into the build pipeline.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.