pdf

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill includes examples that embed plaintext passwords directly into command-line arguments and function calls (e.g., --password=mypassword and writer.encrypt("userpassword","ownerpassword")), which would require the agent to handle or reproduce secret values verbatim.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The “pdf” skill’s form-filling workflow ingests outsider-authored free text directly from user-supplied PDFs and turns extracted/derived text into LLM-relevant inputs at runtime (e.g., extracting form field info and rendering labels/values), as in scripts/extract_form_field_info.py reading PDF annotations/field dictionaries and scripts/fill_pdf_form_with_annotations.py loading user-provided fields.json text.