ClawHub
Back to skill

Security audit

PDF Editor

Security checks across malware telemetry and agentic risk

Overview

This is a coherent PDF tool, but it asks the agent to download and run an unverified proprietary executable before use.

Review before installing if you are comfortable letting the agent download and run ComPDF's proprietary CLI locally. Use extra caution with sensitive PDFs unless you trust that binary and its publisher, and provide the trial-license email only if you accept the activation flow.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Description-Behavior Mismatch

Medium
Confidence
89% confidence
Finding
The manifest markets the skill as a narrow page-editing tool, but the body exposes materially broader capabilities including conversion, optimization, comparison, and watermarking. This scope mismatch can cause the agent or user to invoke more powerful operations than expected, weakening informed consent and review boundaries for a skill that also downloads and runs external code.

Context-Inappropriate Capability

High
Confidence
98% confidence
Finding
The skill requires downloading and executing a binary from a remote server at first use, creating a supply-chain and arbitrary-code-execution risk inside a skill presented as local PDF processing. Even if the domain appears official, the skill provides no integrity verification such as pinned checksums, signatures, or reproducible packaging, so a compromised server, CDN, or network path could deliver malicious code.

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
## Platform Detection (Agent must execute automatically)

Before running any CLI command, the Agent must automatically detect the platform (read `env.platform`: `win32`→Windows, `darwin`→Mac; or run `uname -s`). **Do NOT ask the user about their platform.**


| Variable             | [Windows]                                          | [Mac]                                             |
Confidence
76% confidence
Finding
Do NOT ask the user

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.