chore(docs): refresh blog visual baseline after content change

[wheels-bot]

Summary

A blog-content merge changed the blog index page, so the committed visual-regression baseline at web/tests/visual-baselines/blog.png is stale. Without this refresh, every subsequent web/**-touching PR fails the Visual regression check on a deterministic diff until someone refreshes manually (issue #2915).

develop's branch protection rejects direct pushes, so the refresh lands through this PR, opened and merged by the deploy run that published the blog change.

🤖 Opened automatically by .github/workflows/web-deploy.yml.

[wheels-bot]

Wheels Bot — Reviewer

TL;DR: This PR refreshes the committed visual-regression baseline web/tests/visual-baselines/blog.png after the blog-index drift caused by the "Anatomy of a Wheels Package" post (#3024), exactly as documented in the auto-refresh flow in .github/workflows/web-deploy.yml (issue #2915). The artifact and its provenance check out on every dimension I could verify — verdict: approve (submitted as a comment review because GitHub forbids the wheels-bot identity from approving its own PR).

What was verified

• Artifact validity — both the old and new blog.png are valid 1280 x 800, 8-bit/color RGB PNGs (file on both blobs). The new render shows the blog index with the #3024 post ("Anatomy of a Wheels Package", June 11, 2026) at the top, which is precisely the expected content drift this refresh exists to absorb. No broken assets, debug overlays, or anomalies in the render.
• Provenance — the branch name (docs/bot-refresh-blog-baseline-${GITHUB_RUN_ID}), commit subject, PR title, and PR body all match the generating step at .github/workflows/web-deploy.yml:333-350 verbatim. The PR is only opened when the full visual-regression suite has already passed with the pre-refreshed baseline (steps.regress.outcome == 'success' gate, lines 307-320), so this refresh cannot mask a regression on a non-blog page.
• Commits — chore(docs): refresh blog visual baseline after content change is a valid conventional-commit header (allowed type, ≤ 100 chars, not ALL-CAPS); the Validate Commit Messages check passed. The bot commit carries no DCO sign-off, but this matches the accepted precedent of the prior auto-refresh squash (65188f4eb) and #2678 — noted, not blocking, for a workflow-generated artifact commit.
• Docs / changelog — no fragment needed: a test-baseline binary is not a user-facing change, and chore is the right type.
• Tests — the artifact is the test fixture, and it was validated by the same regression run that produced it.

No correctness, convention, cross-engine, or security findings. (The PR auto-merged at 19:03:39Z via the workflow's poll-and-merge step, consistent with its documented behavior.)