skills/beita6969/scienceclaw/pdf/Gen Agent Trust Hub

pdf

Pass

Audited by Gen Agent Trust Hub on Apr 6, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [DYNAMIC_EXECUTION]: The script scripts/fill_fillable_fields.py performs a monkeypatch on the pypdf library at runtime. This modification of DictionaryObject.get_inherited changes the library's internal logic for attribute retrieval to handle specific PDF structures.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external, untrusted PDF files. Evidence: 1. Ingestion points: User-provided PDF files processed in scripts/ and instructions. 2. Boundary markers: No explicit delimiters or safety warnings are implemented for extracted text. 3. Capability inventory: File system access and shell command execution. 4. Sanitization: Extracted content is processed without sanitization or validation.
  • [COMMAND_EXECUTION]: The skill instructs on and facilitates the use of external CLI tools for PDF manipulation, including qpdf, pdftotext, and pdfimages.
  • [PROMPT_INJECTION]: The forms.md file employs the high-priority keyword 'CRITICAL' to direct agent behavior and workflow compliance, which is a pattern associated with behavioral overrides.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 6, 2026, 11:31 PM
Security Audit — agent-trust-hub — pdf