Session Handoff [review-tooling]: Alternating Sections and the Node Probe Timeout Skip Released as 2.0.1085, Suppressed Read Check Next #3196

Next steps, in priority order

1. Check Suppressed Summaries for a Vetted-but-Unread Section in pr_review.py #3193: check suppressed summaries for a vetted-but-unread section in pr_review.py. unrecognized_in compares read against vetted summaries only for previously-missed sections, so a suppressed section no reader reaches still reports shapes=ok. Done looks like: a constructed body where suppressed_blocks misses a vetted suppressed summary reports it, and the existing previously-missed comparison is extended rather than copied.
2. Finish the Incidental Count Sweep, Including Counted Section Titles #3169: finish the incidental-count sweep. Carried over, lower priority.
3. pr_review.py Passes a Clean-Verdict Copilot Round Whose Headline Sentence Names a Finding #2567: a clean-verdict round whose headline names a finding. Needs a design decision on reading headline prose before any code, so ask before starting it.

External blockers

None.

Internal dependencies

• Check Suppressed Summaries for a Vetted-but-Unread Section in pr_review.py #3193 builds on Read Every Section of a Kind When Sections Alternate in One Wrapper in pr_review.py #3194's walk, which is released.
• pr_review.py Passes a Clean-Verdict Copilot Round Whose Headline Sentence Names a Finding #2567 waits on the maintainer's design call. It is not filed under the decision label, so raise it as a question when it comes up.

State

• Released 2.0.1085 from main at 569af223, the merge of promotion Promote Alternating-Section Reading and the Node Probe Timeout Skip to Main #3195, which carried Tell a Timed-Out Node Probe From an Engine Disagreement in test_spec_validate #3192 and Read Every Section of a Kind When Sections Alternate in One Wrapper in pr_review.py #3194 and closed Read Every Section of a Kind When Sections Alternate in One Wrapper in pr_review.py #3180 and Tell a Timed-Out Node Probe From an Engine Disagreement in test_spec_validate #3182.
• This machine's Codex and opencode skills snapshot was refreshed from 569af223 and reads current. The host-safety kit reports CURRENT.
• The primary checkout is on develop, clean, and fast-forwarded. pr_review.py cleanup removed both task worktrees and branches, locally and on the remote.
• Three detached worktrees prefixed fm- belong to another, live session. cleanup --sweep kept them as too recent. Leave them alone.
• No other repository was touched.

The parked decision queue

The queue held 0 open issues with the decision label when this was written. Nothing is parked.

What the last round did

• Read Every Section of a Kind When Sections Alternate in One Wrapper in pr_review.py #3194 makes marker_blocks visit every bound sibling_heading returns, opening a block only where its marker matches, so alternating sections in one wrapper are each read. Closes Read Every Section of a Kind When Sections Alternate in One Wrapper in pr_review.py #3180 on promotion.
• Tell a Timed-Out Node Probe From an Engine Disagreement in test_spec_validate #3192 makes the node corpus probe in test_spec_validate skip on TimeoutExpired, with a test feeding in both a timeout (it skips) and CalledProcessError (it errors). Closes Tell a Timed-Out Node Probe From an Engine Disagreement in test_spec_validate #3182 on promotion.
• Filed Check Suppressed Summaries for a Vetted-but-Unread Section in pr_review.py #3193 from the local strict review of Read Every Section of a Kind When Sections Alternate in One Wrapper in pr_review.py #3194.
• Promoted through Promote Alternating-Section Reading and the Node Probe Timeout Skip to Main #3195 and released 2.0.1085.

What not to repeat

• A test claiming shapes=ok must build its body from vetted spellings. The first Read Every Section of a Kind When Sections Alternate in One Wrapper in pr_review.py #3180 test used a Markdown ### Previously missed (1) heading, which is unvetted, so on the old code the digest said shapes=UNRECOGNIZED, not the silent shapes=ok the docstring claimed. The local review caught it. Use <summary> spellings from VETTED_SUMMARIES whenever a test is about a silent miss.
• A test that pins a narrow except must also feed it a neighbouring exception. The first Tell a Timed-Out Node Probe From an Engine Disagreement in test_spec_validate #3182 test fed only TimeoutExpired, so widening the handler to SubprocessError would have passed it while turning node refusing a pattern into a skip.
• gh pr merge --match-head-commit takes the full SHA written out. The write guard refuses any argument built at run time.

New learnings

• None beyond the issue filed.