Upload archives, protected files and office documents
Keep the original archive, provide a password privately when requested, and use the files actually read.
Provide common files directly
Upload common spreadsheets, documents, presentations, PDFs, images and text, as well as ZIP, RAR, 7z, TAR and common compressed formats. XLSM, ODS, ODT, ODP, RTF, multipage TIFF and JSON/JSONL/XML are read according to their contents; reading a file does not verify its business accuracy.
State what you need checked. If only handing over files, say “Keep these for now; no analysis.” Uploading does not require generating a report.
Enter passwords privately
When a password is needed, use the private file-unlock prompt in the conversation rather than posting it as chat text. If it does not open the file, check and retry against the same original.
An exported file’s password and a login verification code have different purposes. Follow the current file and platform prompts; do not reuse an old code as a new file password. Automatic SMS assistance depends on the current binding and task.
Handle missing volumes, damage or size limits
One volume of a multipart archive is not the complete file. Provide a complete single archive or extract it locally and upload the needed files. Follow the specific message for damage, unsupported formats, missing volumes or an incorrect password.
Current archive extraction limits are 20,000 entries, 512 MiB in total and 128 MiB per entry. Split larger collections by business period or directory and provide the data needed for the task.
Check what was actually read
Extracted files retain their relationship to the original archive, internal path and source version. Skipped or unreadable parts are not completed work; check the gaps before continuing accounting or analysis.