Security audit

Java代码

Security checks across malware telemetry and agentic risk

Overview

The skill is mostly a Java code review prompt, but it asks for broad command, file-write, external API, and generic API key handling that are not clearly scoped to that purpose.

Review this skill before installing. Use it only in repositories where broad read, command execution, and report-writing access are acceptable, and do not provide a generic API_KEY unless the provider, purpose, transmitted data, and permissions are clearly documented.

SkillSpector

By NVIDIA

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 (5)

Description-Behavior Mismatch

Medium

Confidence: 92% confidence
Finding: The skill is presented as a Java code review generator, but the documentation expands its scope to external API integration and generic system command execution. This capability mismatch increases the risk of users or hosting agents granting broader privileges than necessary, enabling unintended command execution or data exfiltration under a benign-looking skill label.

Context-Inappropriate Capability

Medium

Confidence: 83% confidence
Finding: The skill asks for API key environment configuration even though its stated purpose is local Java code review, and no clear justification is provided. Unnecessary credential handling expands the attack surface by encouraging users to expose secrets to a skill that may not need them, creating opportunities for accidental leakage or misuse.

Intent-Code Divergence

Medium

Confidence: 88% confidence
Finding: The documentation first states that the LLM is provided by the agent, then later instructs users to configure a separate API_KEY, creating ambiguity about which service receives data and credentials. This inconsistency can mislead users into supplying unnecessary secrets and obscures the trust boundary for code and repository contents.

Vague Triggers

Medium

Confidence: 80% confidence
Finding: The activation language is overly broad and overlaps with common development tasks far beyond Java code review, making it easier for the skill to be invoked in unrelated contexts. Overbroad routing is dangerous because a skill with read/write/exec tools may be selected for tasks where the user did not intend to authorize such powerful behavior.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The skill advertises command execution, file writes, and external API use without clear warnings about their system and data impact. In a code review context, these capabilities are unusually sensitive because reviewed source code, diffs, and local files may be modified or transmitted externally without the user appreciating the risk.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.