ClawHub
Back to skill

Security audit

Pdf Processor Cn Tool Free

Security checks across malware telemetry and agentic risk

Overview

The skill is a PDF-processing helper, but it asks for broad write and command execution authority while leaving network callback behavior and file-modifying scope underdefined.

Review this skill before installing. Use it only on PDFs you are comfortable processing locally, avoid callback_url unless you understand the destination and payload, and require confirmation before it installs packages, runs shell commands, writes output files, or deletes temporary files.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (6)

Description-Behavior Mismatch

Medium
Confidence
90% confidence
Finding
The skill is presented as a Chinese PDF processor, but its capability section advertises generic create/query/export/modify/delete operations with no PDF-specific constraints. In a skill that has read/write/exec tools, this mismatch can cause the agent to interpret broad user requests as permission to perform unrelated actions, increasing the risk of unintended file or system operations.

Description-Behavior Mismatch

Medium
Confidence
85% confidence
Finding
A personal-use local PDF processing skill introduces a callback URL and elsewhere references network troubleshooting, implying remote communication outside the stated scope. This creates ambiguity about whether document contents or metadata may be transmitted externally, which is risky for a tool expected to handle potentially sensitive PDFs.

Context-Inappropriate Capability

Medium
Confidence
89% confidence
Finding
The documented remote callback capability is not justified by the stated purpose of a Chinese PDF processing tool and can enable unexpected exfiltration paths. In the context of local document handling, any undocumented outbound destination materially increases privacy and data-leak risk.

Vague Triggers

Medium
Confidence
83% confidence
Finding
The trigger phrases are broad natural-language requests like parsing reports or extracting tables, which can overlap with ordinary conversation and cause accidental activation. Because the skill has exec and write capabilities, unclear activation boundaries increase the chance that the agent performs filesystem or command actions when the user intended only discussion or lightweight assistance.

Vague Triggers

Medium
Confidence
90% confidence
Finding
The capability descriptions repeatedly use generic operational language without scope limits, making it unclear what actions the skill may take and under what conditions. In combination with powerful tools, this ambiguity can lead to overbroad execution, unintended file modifications, or misuse beyond PDF processing.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The manifest grants read, write, and exec capabilities, but the skill description does not warn users that it can modify local files or affect system state. This omission is dangerous because users may reasonably assume a PDF utility only reads documents, while the agent may actually create files or run commands with broader impact.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.