now-reading-page — Claude Code skill: your reading page

What now-reading-page does

The skill spawns a Task subagent that pulls documents from Readwise Reader and writes a standalone HTML file. Data comes in parallel from four locations: shortlist, later, inbox and archive. For the archive the subagent keeps paginating through nextPageCursor until it has roughly 350 documents, enough to cover six months of activity — skip that and the heatmap looks thin.

Results split in two. "Currently Reading" takes items with reading_progress between 0.05 and 0.99, sorted by last_opened_at. "Recently Read" takes archived items above 0.9, grouped by month. There is no "Up Next" section, so the page only shows things you actually opened or finished. The design is warm and sepia-toned: Newsreader and DM Sans from Google Fonts, cover cards with progress bars, a GitHub-style activity heatmap, and JS filter pills for Articles, Books, Tweets, RSS and Email.

Output lands in now-reading/index.html inside the current working directory. You need Readwise access first: either the MCP tools such as mcp__readwise__reader_list_documents, or the readwise CLI with readwise list and readwise read <id>. With neither, there is nothing to fetch — check that before running the skill.