Security audit

PDF Converter

Security checks for vulnerabilities and agentic risk

Overview

This is a scoped PDF conversion helper for ComPDF; it uploads documents only with confirmation and stores only a skill-local API key, though users should understand documents are sent to ComPDF.

Install only if you are comfortable sending selected documents to ComPDF for processing and storing a ComPDF API key in this skill's local api_key file. Keep use to conversion tasks and avoid feeding confidential, regulated, encrypted, or third-party documents into ComPDF unless your organization permits that sharing.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (5)

Description-Behavior Mismatch

High

Confidence: 96% confidence
Finding: The reference file documents a much broader API surface than the skill's declared purpose of general PDF conversion, including decryption, watermark removal, document comparison, PDF generation, and AI extraction. In a skill context, this scope expansion materially increases the chance that an agent or downstream integrator will invoke sensitive capabilities outside user expectations, enabling unintended document manipulation or data extraction.

Context-Inappropriate Capability

Medium

Confidence: 97% confidence
Finding: Exposing PDF decryption in a skill presented as a general conversion utility creates a capability mismatch that can be abused to remove access controls from protected documents. Even if technically legitimate in some enterprise workflows, placing it in this skill lowers the barrier for misuse and increases the risk of unauthorized processing of protected files.

Context-Inappropriate Capability

High

Confidence: 95% confidence
Finding: The AI parsing and extraction endpoints go well beyond file format conversion and enable large-scale content extraction from uploaded documents. In the context of a conversion skill, that hidden expansion is dangerous because it enables data mining of sensitive documents under a less suspicious interface and may expose extracted structured content to third-party processing without informed user intent.

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: The documentation explains direct upload of user files to remote ComPDF infrastructure but does not clearly warn that document contents leave the local environment and are transmitted to a third-party service. For document-conversion workflows, that omission can cause users or agents to send sensitive PDFs, office files, or images to an external processor without informed consent or data-classification checks.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The AI extraction quickstart instructs users to upload source documents for remote parsing and extraction, but omits a clear warning that full document contents and derived structured data may be processed off-platform. This is particularly risky because extraction features can surface sensitive fields at scale, increasing privacy and compliance impact relative to simple format conversion.

Static analysis

No suspicious patterns detected.