pdf-to-epub-ocr

• Outcome: Review
• Latest audit: Aug 7, 2026 · 10:50 AM UTC
• Version: 0.1.0

Overview

This is a coherent local PDF-to-EPUB OCR skill, but it should be reviewed because it pins vulnerable PDF/image-processing dependencies and leaves extracted document artifacts on disk.

Review before installing. Use this only on trusted PDFs unless the dependency pins are updated and resource limits are added. Avoid confidential documents unless you are comfortable with extracted page images, logs, cover files, OCR text inside the EPUB, and conversion outputs remaining in the local output/work directory. Review the generated EPUB against the source PDF because OCR and cleanup may alter content.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (7)

Vague Triggers

Medium

• Confidence: 92% confidence
• Finding: The auto-trigger conditions are broad and ambiguous, covering generic phrases like PDF-to-ebook conversion and any uploaded PDF requesting ebook formatting. This can cause the skill to activate unexpectedly on unrelated or insufficiently scoped user requests, increasing the chance of processing sensitive documents without clear user intent or consent.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The skill does not clearly warn users that uploaded PDFs may contain sensitive information that will be OCR-processed, extracted, and written into output EPUB and report files. In this context, the skill explicitly handles user-supplied documents and generates derivative artifacts, so the lack of disclosure increases privacy and data-handling risk.

Missing User Warnings

Medium

• Confidence: 92% confidence
• Finding: The skill instructs users to provide PDF file paths and describes generating temporary cover images, OCR intermediates, and final EPUB outputs, but it does not clearly warn that uploaded documents and extracted text will be stored and processed on disk. This can expose sensitive document contents or metadata unexpectedly, especially for books or PDFs containing private or copyrighted material.

Missing User Warnings

Medium

• Confidence: 89% confidence
• Finding: The text-cleaning stage removes headers, footers, page numbers, watermarks, whitespace patterns, and performs OCR error correction, but the skill does not warn users that these transformations may delete or alter meaningful content. In structured or technical documents, aggressive cleanup can silently corrupt citations, code, numbering, or legal text, reducing integrity of the converted EPUB.

Natural-Language Policy Violations

Medium

• Confidence: 84% confidence
• Finding: Hard-coding the OCR language to `chi_sim+eng` without user choice or explicit scope constraints can cause significant misrecognition when documents are in other languages or mixed scripts. This is primarily an integrity and reliability issue, but in contexts involving contracts, manuals, or research material, poor OCR can materially distort the output and downstream decisions based on it.

Known Vulnerable Dependency: Pillow==10.1.0 — 10 advisory(ies): CVE-2023-50447 (Arbitrary Code Execution in Pillow); CVE-2024-28219 (Pillow buffer overflow vulnerability); CVE-2026-55379 (Pillow `BdfFontFile`: `Image.new()` called without `_decompression_bomb_check()`) +7 more

Critical

• Category: Supply Chain
• Confidence: 97% confidence
• Finding: Pillow==10.1.0

Known Vulnerable Dependency: PyPDF2==3.0.1 — 2 advisory(ies): CVE-2023-36464 (pypdf and PyPDF2 possible Infinite Loop when a comment isn't followed by a chara); CVE-2023-36464 (pypdf and PyPDF2 possible Infinite Loop when a comment isn't followed by a chara)

High

• Category: Supply Chain
• Confidence: 96% confidence
• Finding: PyPDF2==3.0.1

VirusTotal

VirusTotal findings are pending for this skill version.

Static analysis

No suspicious patterns detected.