Security audit

Csv Insight Free

Security checks for vulnerabilities and agentic risk

Overview

This appears to be a local CSV analysis skill with some documentation and scoping issues, but no evidence of hidden, destructive, or data-exfiltrating behavior.

Install only if you intend to use it for explicit local CSV analysis. Review output paths before export, avoid running it on sensitive CSVs unless you are comfortable with derived files being written to disk, and be aware that the package currently lacks the actual script referenced by the examples.

A.I.G

Vulnerability Patterns

Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Behavioral ASTexec() Call, eval() Call, Dynamic Import
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

Findings (6)

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: The skill advertises itself for very broad, common activities such as data analysis, report generation, and visualization, which increases the chance an agent will invoke it in contexts where file execution or local data handling was not specifically intended. Because the skill declares read, write, and exec capabilities, overbroad invocation cues can cause unnecessary exposure of local files and command execution in routine workflows.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 92% confidence
Finding: The documentation promotes export/save behavior but does not clearly warn that these actions write files and may overwrite or persist potentially sensitive user data. In the context of a skill with write and exec permissions, omission of persistence and overwrite warnings can lead to unintentional data disclosure, residue on disk, or destructive file replacement.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 89% confidence
Finding: The capability coverage section uses expansive keyword-based activation cues, including highly generic terms, to justify applicability across many prompts. In an agent ecosystem, this can lead to over-triggering of a skill with exec and write permissions, increasing the risk of unintended file processing, persistence of user data, and unnecessary command execution.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 97% confidence
Finding: L274 将“LLM API”标记为“必需”，意味着该技能运行依赖外部 LLM 服务；但 L280-L281 又明确写明“基于 Python 标准库，无需额外 API Key”“无需安装任何第三方包”。这不是单纯信息不全，而是对运行依赖的相互冲突描述，会误导用户对技能实际依赖面的理解。
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Low

Category: Not specified by scanner
Confidence: 96% confidence
Finding: L007 在 front matter 中将许可证写为 Proprietary，但 L309 又明确写“改进 license：MIT”。许可证是技能意图与分发条件的重要文档信息，这两处表述互相冲突，属于文档与文档间的明确不一致。
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

Low

Category: Not specified by scanner
Confidence: 76% confidence
Finding: Manifest 和主体描述都将该技能定位为本地、零外部依赖的 CSV 分析工具，不适用于流式处理，且核心能力集中在统计、筛选、异常检测和导出。L324 却把“网络错误”“ping 命令测试网络连通性、防火墙和代理设置”写入技能错误处理，这为技能引入了与其本地 CSV 分析定位不一致的网络运维语境。
Content: No source excerpt is available for this finding.