Security audit

uos-cve-checker

Security checks for vulnerabilities and agentic risk

Overview

This skill is a UOS CVE lookup helper that uses the network and writes local result files in ways that match its stated purpose.

Install only if you want an agent to query UnionTech security advisory services for UOS CVE information. Run it from a working directory where overwriting cve_info.csv is acceptable, and use -d only when you intentionally want to download patch files from the advisory-provided links.

SkillSpector

Vulnerability Patterns

MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
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

Findings (5)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 94% confidence
Finding: The skill advertises and instructs use of capabilities that read input files, write result files, and perform network access, but it does not declare permissions or provide an explicit capability boundary. This creates a transparency and consent problem: an agent or user may invoke a skill that changes local state and reaches external services without clear prior disclosure, increasing the risk of unintended file access, data creation, or network egress.

Description-Behavior Mismatch

Medium

Confidence: 88% confidence
Finding: The skill description says it checks CVEs and obtains fix information, but when the -d flag is used it also downloads patch files and writes them to disk. This mismatch can mislead users or higher-level agents into granting broader side effects than expected, creating integrity and storage risks from unreviewed remote content.

Description-Behavior Mismatch

Medium

Confidence: 91% confidence
Finding: The script writes cve_info.csv and may also create architecture-named directories and patch files, but these local write side effects are not reflected in the skill description. Undisclosed file creation can surprise users, overwrite expected artifacts, and violate sandbox or least-privilege assumptions in agent environments.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The markdown states that the skill can automatically download patches and generate result files, but it does not prominently warn the user that invoking the skill may modify the local filesystem and retrieve content from the network. In an agent setting, this can lead to unexpected state changes, storage of untrusted files, and accidental execution flow toward downloading security patches without informed consent.

Missing User Warnings

Low

Confidence: 82% confidence
Finding: The script unconditionally creates or overwrites cve_info.csv without an explicit warning or confirmation. In automated environments this can destroy prior output, pollute working directories, or create misleading trust in generated files.

Static analysis

No suspicious patterns detected.