The Agent Skills Directory

[SAFE]: The skill uses the docling library, which is a reputable document understanding framework from IBM Research.
[PROMPT_INJECTION]: The skill provides an attack surface for Indirect Prompt Injection.
Ingestion points: Untrusted document files (PDFs, Word documents, and Images) are processed via DocumentConverter().convert() as described in SKILL.md.
Boundary markers: Absent. There are no instructions or delimiters designed to isolate the extracted content from the agent's control flow.
Capability inventory: The code examples in SKILL.md include file system write operations such as os.makedirs, element.image.save, and writing results to disk using open().write.
Sanitization: Absent. No filtering or validation of the extracted document text is performed before it is returned to the agent context.

doc-parser