ClawHub
Back to skill

Security audit

pdf

Security checks across malware telemetry and agentic risk

Overview

This PDF skill does local PDF processing as advertised, but users should handle decrypted PDFs and generated form files carefully.

Install if you need local PDF processing. Use it on copies of important files, keep generated JSON/PNG/PDF outputs in a controlled folder, delete intermediates when finished, and only decrypt PDFs when you have permission to access and store the unprotected result.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (3)

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The skill includes a qpdf example for removing PDF passwords and decrypting protected documents without any warning about authorization, legality, or handling of sensitive content. In an agent setting, this can normalize bypassing document protections and lead users or downstream agents to process confidential files without verifying they are permitted to do so.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The workflow explicitly instructs the user to generate field metadata, page images, JSON mappings, and a filled PDF, but it provides no warning that these artifacts persist sensitive personal data on disk. In a PDF-form-processing skill, those intermediate files can expose PII or overwrite existing files if written into shared or predictable locations, increasing the chance of accidental disclosure.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The non-fillable workflow requires creating validation images and annotated PDFs that visually embed form contents, which may include highly sensitive personal or regulated information. Because the instructions omit any data-handling precautions, the process can leak sensitive data through retained images, JSON files, or derivative PDFs, especially in shared workspaces or synced directories.

VirusTotal

67/67 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.