dsh-attachment-formats

LIMITATIONS

Known limitations

- OCR (tesseract.js) quality is limited on low-resolution scans and complex tables; insufficient confidence falls back to page images with an explicit note — garbled text is never injected. Higher-quality OCR (RapidOCR/MinerU/PaddleOCR) can be added as pluggable backends later (see `docs/upgrade-v6.md`). - The pymupdf4llm high-fidelity engine handles ≤40-page PDFs only (larger documents use the fast pdfjs engine); table/formula reconstruction is good but not typesetting-grade — layout details can be cross-checked against page images. - Scanned PDFs without a text layer can only go the page-image route when OCR is unavailable or fails (vision models can read them). - Legacy `.doc/.xls/.ppt` require LibreOffice (`soffice`); `rtf` requires pandoc; `epub/odt` work out of the box but pandoc (if installed) gives better fidelity. Missing binaries produce clear, actionable errors — nothing is silently dropped. - DOCX formulas and embedded images are not extracted (tables, headings and text are). - XLSX outputs displayed text/results only; charts and comments are not extracted. - Outlines prefer bookmark TOCs; PDFs without bookmarks fall back to font-size heuristics (weak on documents without strong heading styling) — the index card still carries line/page counts and reading pointers. - iWork and archives are not converted yet. - Attachments are attributed to the shell's **current conversation** (the one being viewed). Text/document cards therefore land in the dialog you are looking at. Converted page images go through the harness's native drop pipeline: if the current conversation is mid-reply it temporarily refuses drops, so the plugin waits for it to become idle before feeding the images. With several conversations open at once, other *idle* conversations may also accept that same synthetic drop — a harness-level behavior the plugin cannot scope; prefer attaching images with a single conversation open (text/code files are unaffected: they always stay in the current dialog). - The "merge on send" for document cards bridges into the React controlled input over DOM events — an adaptation to an unpublished harness API; if a core upgrade breaks it, the symptom is "card content didn't enter the message", and the card's **send** button is the fallback (synthetic Enter path). The image path is never affected.