The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses several command-line tools for processing PDF documents, including pdftotext, qpdf, pdftk, pdftoppm, and pdfimages.
[PROMPT_INJECTION]: The skill has an indirect prompt injection surface because it extracts and processes text from untrusted PDF documents using libraries like pypdf and pdfplumber.
Ingestion points: SKILL.md, scripts/check_fillable_fields.py, scripts/extract_form_field_info.py, scripts/fill_fillable_fields.py, and scripts/fill_pdf_form_with_annotations.py involve reading external PDF files.
Boundary markers: No specific delimiters or warnings are used to separate extracted data from agent instructions.
Capability inventory: The skill possesses file write permissions, shell execution capabilities, and image conversion tools.
Sanitization: There is no evidence of sanitization or filtering applied to the extracted PDF text.
[PROMPT_INJECTION]: Procedural instructions in forms.md utilize imperative language ("CRITICAL: You MUST complete these steps in order") to enforce a specific workflow for form-filling tasks.
[SAFE]: The script scripts/fill_fillable_fields.py implements a runtime monkeypatch of pypdf.generic.DictionaryObject.get_inherited. This is a documented technical workaround to fix a specific bug in the library's handling of selection list fields and does not represent a security threat.

pdf