Security audit

CSV文件处理(免费版)

Security checks for vulnerabilities and agentic risk

Overview

This CSV skill is mostly a local file-processing guide, but it asks for broad command execution and mentions API keys/external API integration without a clear CSV-related need.

Review this skill before installing. It may be acceptable if you only use it on local CSV files and your agent asks before running shell commands, but avoid giving it API keys or allowing unrestricted command execution unless you have a clear reason and trust the workflow.

SkillSpector

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
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 (2)

Context-Inappropriate Capability

High

Confidence: 89% confidence
Finding: The documentation claims 'API integration' and external service response handling for a CSV utility, despite the described purpose being local CSV processing. In an agent ecosystem, overstating network-facing capability can cause the skill to be invoked in broader contexts and may justify unnecessary expansion of trust boundaries or future tool access, increasing the risk of data exfiltration or misuse.

Context-Inappropriate Capability

High

Confidence: 95% confidence
Finding: This CSV-processing skill declares and promotes command execution capability beyond what is justified by its stated purpose, and the manifest allows the exec tool. In agent environments, unnecessary command execution materially increases attack surface: malicious file names, shell interpolation, or future implementation drift could turn a simple data-processing skill into arbitrary code execution or local data access.

Static analysis

No suspicious patterns detected.