Security audit

xlsx-data-wizard

Security checks for vulnerabilities and agentic risk

Overview

This Excel skill is mostly a local workbook-processing guide, but it contains contradictory API credential instructions and underspecified destructive spreadsheet operations that users should review before installing.

Install only if you are comfortable supervising file writes. Do not provide API keys or credentials based on this skill alone; the inspected artifact does not clearly define any legitimate external API integration. For important spreadsheets, work on copies and require explicit confirmation before overwriting files or deleting sheets.

SkillSpector

Vulnerability Patterns

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
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (4)

Intent-Code Divergence

Medium

Confidence: 96% confidence
Finding: The skill documentation gives conflicting guidance about whether API keys are required, which can mislead an agent or operator into sourcing credentials unnecessarily or handling secrets in the wrong workflow. In a skill that processes potentially sensitive Excel data, ambiguity around credential usage increases the risk of unintended secret collection, misconfiguration, and trust in undocumented external connectivity.

Intent-Code Divergence

Medium

Confidence: 95% confidence
Finding: The skill first states that online spreadsheet/API integrations are out of scope, then later instructs users to establish API connections and call interfaces. This contradiction can cause an agent to perform networked actions the user would not expect from a local Excel-processing skill, potentially exposing workbook contents or prompting for credentials under false assumptions.

Intent-Code Divergence

Low

Confidence: 94% confidence
Finding: The FAQ says API keys are required for some features, which conflicts with the earlier statement that no extra API key configuration is needed. Even if not directly exploitable on its own, this inconsistency can trick operators into entering secrets unnecessarily or make downstream tooling assume secret-dependent behavior that the skill should not need.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The skill instructs the agent to write, modify, delete, rename, and save Excel content without clearly warning about overwriting originals or making destructive changes. In document-processing workflows, this can lead to silent data loss, corruption of important spreadsheets, or modification of evidence-bearing financial and inventory records.

Static analysis

No suspicious patterns detected.