skills/kursku/skills/pdf-official/Gen Agent Trust Hub

pdf-official

Pass

Audited by Gen Agent Trust Hub on May 2, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external PDF files using various libraries (e.g., pypdf, pdfplumber). This creates a surface for indirect prompt injection if an attacker embeds malicious instructions within a PDF's text or metadata. This is an inherent risk of any document processing skill, and the skill provides structured workflows to mitigate errors.
  • Ingestion points: SKILL.md (Quick Start), forms.md (Visual Analysis), and scripts like extract_form_field_info.py.
  • Capability inventory: File system read/write operations for PDF/PNG/JSON files, and execution of local Python scripts.
  • Sanitization: Not explicitly implemented in the provided code snippets, though the scripts use standard library parsers.
  • [DYNAMIC_EXECUTION]: The script scripts/fill_fillable_fields.py implements a runtime monkeypatch of the pypdf library's DictionaryObject.get_inherited method. This is used as a transparent workaround for a specific formatting bug in the library when handling selection lists. The logic is clearly documented and scoped strictly to fixing data types for form fields.
Audit Metadata
Risk Level
SAFE
Analyzed
May 2, 2026, 05:49 PM
Security Audit — agent-trust-hub — pdf-official