Re-audit Blog against the hugo type and bump the conformance matrix

[ptr727]

The 2026-08-03 report graded Blog as source-only against a hub that had no static-site type, and it recorded that it was due a re-run once the deploy existed. This is that run, replacing the file in whole rather than editing it, per AUDIT.md section 8.

Run stamp audit run 2026-08-05T21:57:38Z | hub 01507a0, read at Blog main@2b132e4. Verdict operational, no defect.

Judged for the First Time

All nine hugo.* checks, each cited to file:line. They pass. The three deviations the first run recorded are closed:

Issue	What it was	What closed it
ptr727/Blog#27	The remote release tree was never pruned	A host timer owns the prune, recorded at OPERATIONS.md:170 with its count and the reason the deploy path does not prune
ptr727/Blog#28	The vendored theme recorded no upstream ref	themes/README.md:10-16 records the repository, commit, date, describe form, and license, kept outside the vendored directory so an update does not take the record with it
ptr727/Blog#29	The generator pin was duplicated across two workflows	One composite action both callers use, with no overridable input, so agreement is structural rather than asserted

hugo.deploy.retention takes the second of the two D5.6 shapes, which is the correct one here: the deploy credential is confined write-only and cannot observe the destination, so an in-pipeline assertion would mean widening it. Note that deploy/make-release.sh does prune, but in CI it runs against a scratch bundle path, so it satisfies nothing on its own and the host timer is what the verdict rests on.

Develop drift is benign: behind 6, ahead 0, identical trees.

Two Drift Classes Stay Open

Both are the hub having advanced past what the repo carries, not anything the repo did. Seven stale verbatim units and 17 prose findings, each read rather than counted.

Three of the seven differ from the canonical only in the case of the word Markdown (#566). The other four are rule additions the repo states in an older form, from #569, #561, #560, and #563.

One needs sequencing. The issue-closing-keyword rule moved out of Release Model into Branching Model in #563. A re-vendor that takes the new Release Model without also taking the new Branching Model drops the rule entirely rather than leaving it stale.

A Hub Defect the Run Surfaced

spec/audit.py lines 691-695 flag a carried file that names the hub anywhere in it. The check has no exemption for AGENTS.md > Fleet Bootstrap, which is byte-locked and whose first sentence names the hub by path, because saying where the canonical rules live is that section's entire function. Blog's line 11 is byte-identical to this repo's own AGENTS.md:11 and is the file's only occurrence.

The finding is unclearable by construction: satisfying the check means altering a section the verbatim check would then flag. It is not specific to Blog. It fires on any repo that has carried the current canonical, so it spreads as the carry propagates rather than staying a single-repo curiosity. Recorded as escalation 1 with the fix, which is to scan the file with the Fleet Bootstrap block excised so a reference outside it is still caught. Not counted against Blog's verdict, and not fixed here, since it is a change to the runner rather than to a report.

Two further escalations: a downstream repo holding a report at the hub's own reports/<repo>/audit.md path, which section 8 does not address, and the unclassified COPYRIGHT SIGN carried unchanged from the first run.

The Matrix

The hugo row takes this run's date and findings. Ten clause-joining semicolons and three spaced hyphens elsewhere in the file are fixed in the same pass rather than stepped around.

The 17 remaining dash findings are deliberately left. Every one is a | - | empty-table-cell placeholder, the notation the file's own header documents, so they are a prose_lint.py exemption gap rather than divergent content. The gap is systemic, 46 such cells across 8 report files, and a sweep to satisfy the checker would damage every one of them. Two of the semicolons fixed here are ones the checker missed, one inside parentheses and one suppressed by the colon-in-row exemption, so the exemption cuts both ways. Reported rather than patched, because changing the checker needs its own change with an old-versus-new verdict diff.

Verification

prose_lint.py clean on the report and on every line changed in the matrix. markdownlint-cli2 clean over reports/**/*.md. editorconfig-checker clean over the tree, after converting the new report to CRLF to match the [*] default. spec/validate.py reports 22 cataloged and 0 backlog repos classifying cleanly. Reference-link integrity checked in both files: no undefined references, no unused definitions, no inline targets, and every local path resolves.

Documentation only. No behavior change, and no registry edit: the entry already declares both types, both publish targets, and four driftNotes that all describe live deviations.

🤖 Generated with Claude Code

[Copilot]

Pull request overview

Updates the hub’s documentation evidence for the Blog repository by replacing its audit report with a 2026-08-05 re-run (now evaluated against the hugo type) and refreshes the conformance matrix row to reflect that run.

Changes:

• Replace reports/blog/audit.md with the 2026-08-05 audit run, including per-check Hugo evidence and updated drift/escalation notes.
• Update the hugo row in reports/conformance-matrix.md with the new audit date and updated notes, plus small prose tweaks in the matrix header text.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 2 comments.

File	Description
reports/conformance-matrix.md	Updates the Blog (hugo) row audit date/notes and tightens explanatory prose for shapes/targets.
reports/blog/audit.md	Replaces the Blog audit report with the 2026-08-05 run, expanding Hugo-dimension evidence and revising drift/escalations accordingly.

[Copilot]

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated no new comments.