skills/wildwasser/opencode-agents/pdf/Gen Agent Trust Hub

pdf

Warn

Audited by Gen Agent Trust Hub on May 27, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The script scripts/fill_fillable_fields.py implements a runtime modification of the pypdf library's internal logic.
  • Evidence: The monkeypatch_pydpf_method function reassigns the get_inherited method of pypdf.generic.DictionaryObject to a local wrapper function at runtime. This dynamic execution technique is used to bypass a bug in the library's handling of selection list fields.
  • [PROMPT_INJECTION]: The skill's primary function involves ingesting and processing untrusted data from PDF files, creating a vulnerability surface for indirect prompt injection.
  • Ingestion points: Data is read from external PDFs via pypdf, pdfplumber, and pdftotext across multiple scripts and instructions.
  • Boundary markers: The instructions lack specific delimiters or instructions to the agent to ignore or isolate potentially malicious commands embedded within the extracted PDF text.
  • Capability inventory: The skill provides the agent with file system access (read/write) and instructions to execute shell commands based on the interpreted content of the documents.
  • Sanitization: There is no logic present to sanitize or validate extracted text or field metadata before it is used in the agent's context or further processing steps.
  • [COMMAND_EXECUTION]: The skill documents and encourages the execution of various command-line utilities for PDF processing.
  • Evidence: SKILL.md and reference.md contain specific usage instructions and examples for executing qpdf, pdftotext, pdfimages, and pdftk using shell subprocesses.
Audit Metadata
Risk Level
MEDIUM
Analyzed
May 27, 2026, 02:35 PM
Security Audit — agent-trust-hub — pdf