firecrawl/anydoc

The `anydoc` XLSX conversion silently exposes hidden rows and columns as visible content, both in its structured document model and Markdown output. This lack of visibility metadata forces developers to process potentially irrelevant or sensitive internal data, such as calculation or template content, as if it were user-visible. The absence of an option to filter or flag hidden content increases the risk of feeding misleading or unintended information into downstream systems, particularly LLMs, where such data could be misinterpreted as authoritative. This behavior undermines data accuracy and control. For B2B SaaS solutions relying on `anydoc` for document ingestion, this flaw introduces significant data quality and security risks. Exposing hidden data without context can lead to erroneous analyses, compliance issues, or the inadvertent leakage of proprietary information. Enterprises require precise control over data visibility during conversion. A tool that flattens visibility without options for control is less suitable for applications demanding high data fidelity and contextual awareness, potentially limiting `anydoc`'s adoption in regulated or data-sensitive environments.