ClawHub
Back to skill

Security audit

图纸解析

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed drawing/PDF/CAD parser that writes local extraction results, with no evidence of hidden data access, exfiltration, persistence, or destructive behavior.

Install this only for technical drawing parsing and prefer invoking it with explicit CAD/PDF/building-drawing language. For safer production use, pin dependency versions and review optional OCR/CAD tools before installing them.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (6)

Vague Triggers

Medium
Confidence
90% confidence
Finding
The trigger keyword "读图" is very broad and can match ordinary user requests about reading or interpreting images, diagrams, or documents beyond the intended CAD/PDF parsing scope. In a tool skill, overly broad invocation can cause unintended activation on unrelated content, increasing the chance of unnecessary file handling, privacy exposure, or execution of parsing workflows on user data without clear intent.

Unpinned Dependencies

Low
Category
Supply Chain
Content
# 基础依赖（必需）
pymupdf>=1.23.0
ezdxf>=1.4.0
openpyxl>=3.1.0
numpy>=1.24.0
Confidence
89% confidence
Finding
Using a lower-bound specifier like 'pymupdf>=1.23.0' allows installation of any newer release, including versions with breaking changes or newly introduced malicious/compromised transitive content. In a document-parsing skill that processes untrusted PDFs/CAD files, uncontrolled dependency drift increases supply-chain risk and can expose the parser to vulnerable library versions.

Unpinned Dependencies

Low
Category
Supply Chain
Content
# 基础依赖（必需）
pymupdf>=1.23.0
ezdxf>=1.4.0
openpyxl>=3.1.0
numpy>=1.24.0
matplotlib>=3.7.0
Confidence
89% confidence
Finding
Using 'ezdxf>=1.4.0' leaves dependency resolution open-ended, which can pull in unreviewed future releases. Because this skill parses CAD drawings from potentially untrusted sources, dependency instability or a compromised upstream package could directly affect a high-risk file-processing path.

Unpinned Dependencies

Low
Category
Supply Chain
Content
# 基础依赖（必需）
pymupdf>=1.23.0
ezdxf>=1.4.0
openpyxl>=3.1.0
numpy>=1.24.0
matplotlib>=3.7.0
Confidence
86% confidence
Finding
An unpinned 'openpyxl>=3.1.0' can result in nondeterministic environments and unexpected exposure to future package defects. While this library is not the primary parser here, it still participates in the processing pipeline and contributes to supply-chain attack surface.

Unpinned Dependencies

Low
Category
Supply Chain
Content
pymupdf>=1.23.0
ezdxf>=1.4.0
openpyxl>=3.1.0
numpy>=1.24.0
matplotlib>=3.7.0

# OCR 模式选择（按需安装）
Confidence
84% confidence
Finding
Specifying 'numpy>=1.24.0' permits any later version, reducing build reproducibility and increasing exposure to supply-chain or compatibility issues. Although numpy is a common dependency, leaving it unpinned still weakens control over the runtime environment.

Unpinned Dependencies

Low
Category
Supply Chain
Content
ezdxf>=1.4.0
openpyxl>=3.1.0
numpy>=1.24.0
matplotlib>=3.7.0

# OCR 模式选择（按需安装）
# rapidocr        # v6模式（推荐，PP-OCRv6）
Confidence
83% confidence
Finding
Using 'matplotlib>=3.7.0' allows future releases to be installed without review, which can introduce unexpected code paths, vulnerabilities, or operational failures. In this skill it is likely ancillary, so the direct security impact is lower, but it still contributes to avoidable supply-chain risk.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.