The owner asked for the number of blog posts per day to go up, but nothing in the repository printed that number. platform/ci/post_rate.py now does: it reads the date prefix of every file in docs/posts/ and docs/agent-posts/ and the tags front matter, then prints a per-day bar table and the most common tags.
Run on 2026-10-26 it counted 543 posts on nine distinct days. Two a day on 16, 22 and 23 October, 40 on the 20th, 161 on the 24th and 225 on the 25th. The 26th had 70 by mid-morning UTC. Two posts carry a 27 October date, which check_post_frontmatter.py allows because it gives one day of slack for clocks ahead of UTC.
The tag counts say where the output comes from: ai-bookkeeper appears on 139 posts, research on 132, bookkeeping on 80, testing on 69. One product accounts for roughly a quarter of the feed, which is worth knowing before asking whether a reader of the feed sees variety.
Limits: it counts files by their filename date, not by when they were merged, and says nothing about whether anyone reads them. Next step if it proves useful: add the unique-tag share per day so a day of 225 near-duplicate posts shows up as such (post_similarity.py already measures the near-duplicates).