skills/thienqo/ios-skills/pdfkit/Gen Agent Trust Hub

pdfkit

Pass

Audited by Gen Agent Trust Hub on Oct 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process and display external PDF documents which constitutes an attack surface for indirect prompt injection.
  • Ingestion points: The skill uses PDFDocument(url:) and PDFDocument(data:) in references/core-implementation.md to load content from external sources.
  • Capability inventory: In references/pdf-viewing-and-annotations.md, the ManagedPDFView coordinator implements a link handling delegate that calls UIApplication.shared.open(url). This allows the agent/application to navigate to arbitrary external websites if a link within a PDF is triggered. Additionally, the skill provides capabilities for file system writes via document.write(to:).
  • Boundary markers: There are no specific instructions or delimiters provided to the agent to treat embedded PDF content or links as untrusted.
  • Sanitization: The provided implementation does not include logic for sanitizing or validating URLs before they are passed to the system's URL opener.
  • [EXTERNAL_DOWNLOADS]: The skill references documentation hosted on a non-standard domain.
  • Evidence: Multiple links in SKILL.md point to https://sosumi.ai/documentation/pdfkit and related paths. While presented as documentation for the framework, this domain is not an official Apple developer resource and is not part of the trusted vendor list.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 9, 2026, 12:02 PM