This is a real handoff, written partway through work on this blog series: the session that adopted the six-post structure, wrote the introductory post, and picked the HTML template every post since has used. Post titles, file paths, and a few other specifics have been generalised so the example reads on its own instead of leaning on this project's internal names. The structure, the caps, the labelled fields, the terse noun phrases, is exactly how it came out.
HANDOFF v3.0 | 2026-09-09 | supersedes: prior blog-series handoff (same date) | session-length: long
OBJ: Rework two existing posts to a new six-post editorial blueprint, write a new introductory post, revise it through a full review-comment pass with the user, design and pick an HTML template for the series, and commit everything to the user's machine.
STATE
content/post-0.md | new file | ~740 words total (659 excl. appendix) | committed, through 4 rounds of user review comments plus 1 post-decision typo fix
content/post-0-template-c.html | new file | winning HTML template, chosen by the user over 2 alternatives | committed, this is the template to reuse for the rest of the series
content/post-0-template-a.html, content/post-0-template-b.html | new files | the two HTML variants not chosen | kept in the repo for reference, not deleted
content/post-1.md | overwritten | 2,211 words total (1,542 excl. appendix) | committed | not yet HTML-exported or user-reviewed
content/post-2.md | overwritten | 1,920 words total (1,412 excl. appendix) | committed | not yet HTML-exported or user-reviewed
content/ (older draft files) | untouched, an earlier abandoned two-post draft, left in place
DECIDED
Adopt the six-post blueprint fully, supersedes the prior "five posts, one per skill, no separate foundations post" plan
Post 0's title and subtitle finalised, after weighing two shorter alternatives
Post 1's title finalised, replacing an earlier working title flagged for corrective-antithesis phrasing
Post 2's title finalised, replacing an earlier working title flagged for the same reason
British spelling ("judgement") used consistently across all three posts, matching the user's existing global preference
User explicitly overrode two of the project's own banned-phrase examples for Post 0's opening paragraph, by direct instruction, after an earlier draft tried to route around them
Post 0's word count (659 excl. appendix) is now well under the blueprint's 900-1,300 target, a direct and accepted result of the user's trim-heavy edits; flagged twice, user did not ask for it to be padded back up
HTML styling must be derived from the user's own existing site, not invented generically: serif headings, sans body, warm cream / dark palette, manual light-dark toggle plus a font-size toggle, hairline borders, card and pill conventions carried over from the user's own site patterns
Of 3 HTML template directions offered (A: card TL;DR with circular badges; B: editorial pull-quote with drop-cap; C: compact monospace technical/documentation style), the user picked C. C is now the template standard for the rest of the series
Post 0 delivered as markdown (source of truth) plus 3 HTML variants; the user plans to wire the chosen one into the live site later, not this session
PROPOSED
Apply the same comment-review pass and the chosen HTML template to the other two existing posts, once the user is ready
REJECTED
Reusing an old filename for the new Post 0 | reason: that name already belonged to an older, abandoned draft on disk
Keeping an expanded, multi-paragraph "what changed" section | reason: the user found the expanded AI-drafted prose "bland word vomit" and cut it back down through direct edits
Generic, invented HTML styling for the blog pages | reason: the user already has a real, designed site; the first HTML attempt ignored it and had to be redone from the site's own markup
HTML Version A (card TL;DR) and Version B (editorial pull-quote) | reason: the user picked Version C (compact technical/monospace) instead
OPEN
The remaining three posts in the series are not yet started against the blueprint | blocks: series completion
Whether an older draft's content is salvageable for the last two posts | blocks: drafting those posts
Whether the other two existing posts should go through the same comment-review and HTML-export pass as Post 0 | blocks: nothing yet, the user hasn't asked
UNVERIFIED
Whether three skill proposals shown earlier in the session were actually saved by the user | not re-checked this turn
PATTERN
The user strongly prefers direct, minimally-processed prose over AI-smoothed expansions; multiple rounds this session where added connective or explanatory sentences were rejected in favour of the user's own shorter original phrasing, even when that meant reintroducing wording the project's own style guide names as something to avoid
Before producing any visual or design deliverable, check the user's existing real assets first rather than inventing a generic version; the first HTML draft skipped this and had to be redone
FIRST
Ask the user whether they want to run the other two existing posts through the same review-comment pass as Post 0 next, using the now-chosen template, or move on to drafting the next post in the series first.A few things worth noticing that aren't obvious from reading it once:
- REJECTED isn't a graveyard of abandoned ideas. Each line names a reason, so a later session doesn't re-propose a direction that was already tried and turned down.
- PATTERN holds two lines here, and each one recurred more than once across the session before it earned a place. A single moment of friction doesn't qualify; only a repeat does.
- FIRST is one line, one action. Everything else in OPEN can wait for a decision; FIRST is what actually happens the moment the next session begins.