Security audit

Spreadsheet & Data Wrangling Master

Security checks for vulnerabilities and agentic risk

Overview

This is a spreadsheet guidance skill with expected data-cleaning and reporting instructions, not hidden code or deceptive behavior.

Install is reasonable for spreadsheet and CSV work. Use it on copies of important business or financial files, ask the agent to preview cleaning actions first, and require confirmation before overwriting files, deleting rows, deduplicating records, or changing source spreadsheets.

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
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 (3)

Vague Triggers

Medium

Confidence: 94% confidence
Finding: The quick-start phrases are generic natural-language commands like 'Clean this CSV' and 'Analyze this data' that can easily overlap with ordinary user requests. In systems that auto-activate skills based on README trigger text, this can cause unintended invocation of the spreadsheet skill, potentially changing agent behavior, exposing data to the wrong processing workflow, or bypassing a more appropriate specialized skill.

Vague Triggers

Medium

Confidence: 94% confidence
Finding: The skill exposes very broad natural-language triggers such as 'Analyze this data', 'Build a dashboard', and 'Clean this CSV' that closely match common user requests. In an agent environment that auto-selects skills by phrase matching, this can cause unintended invocation and execution of file-modifying workflows without the user deliberately choosing this skill.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The skill advertises operations like cleaning CSVs, converting files, creating reports, and making data tidy, all of which imply transforming or creating artifacts, but it does not prominently warn that these actions may change user data or produce derivative files. In practice, this increases the risk of accidental data loss, silent overwrites, or unintended modifications when the agent applies the skill automatically or without a confirmation step.

Static analysis

No suspicious patterns detected.