The Agent Skills Directory

[SAFE]: The skill implements PDF processing using established libraries and follows a structured workflow with local scripts for form handling. No indicators of malicious activity, data exfiltration, or harmful command execution were found.
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection because it extracts content from untrusted PDF documents. Ingestion points: Text extraction snippets in SKILL.md and form field extraction in scripts/extract_form_field_info.py. Boundary markers: None explicitly defined in the provided instructions for processing extracted data. Capability inventory: File writing and script execution across the scripts/ directory. Sanitization: No evidence of content sanitization prior to presentation to the agent.
[SAFE]: The script scripts/fill_fillable_fields.py performs a monkeypatch on the pypdf library to resolve a specific known issue with selection list formatting. This is an intended functionality for accurate form filling.

pdf