[FBumann]

One section in CONTRIBUTING.md, 14 lines. No code, no gate, no tool.

Kept to rules on purpose. The file's own charter is "readable in one sitting
and go stale only when a command changes", so the evidence below stays here in
the PR rather than in the page — which is the same split docs/ already runs on.

The failure the rule is for

Triaging the backlog surfaced the same defect four times, and it is mechanical
rather than a matter of care:

issue	rested on	outcome
#50	_sum_piece / _group_piece / _shift_piece at executor.py:486-539	closed, re-filed as #332
#103	a _group_fragment SQL fragment	closed, re-filed as #431
#27	"grep raise RelationalBuildError in lowering.py"	closed — neither the class nor the module exists
#244	bench/results/latest.jsonl on the polars-engine branch	branch deleted; needs a re-measure

In every case the finding survived the rewrite and the evidence for it
did not, so the issue had to be closed and re-filed rather than read.

#474 then found the same rot pointing the other way — code and tests linking to
ROADMAP sections (Track 5, ROADMAP 2c, ROADMAP 5c, "The degree axis")
that never existed or no longer did, two of them inside user-facing error
messages. Same rule, both directions.

The order half

• now, capped at five — currently A scalar variable cannot be masked: presence frames have no column to hold the empty coordinate #340, The where grammar takes no quoted string and no datetime, and checks no dtype #460, The breaking-surface window: sequence #67, #332 and #29 as one decision #432, at(): a consumer report — two downstream features blocked, and what the surface needs to cover #185, optional: on parameter declarations — an absent table means no rows #217. The cap is
  the mechanism: a claimed priority order over forty-six issues is fiction, and
  five is the smallest thing that is not.
• Grouping is sub-issue parentage, not a label. Tracks are Track 1 — primitives #470, Track 2 — the operational surface #471 and
  Track 3 — capabilities, and the degree line #472 with 21 sub-issues between them.
• blocked:decision, the sibling blocked:upstream never had — Namespacing: qualified names for component libraries #29, Shape operators: pushforward/pullback taxonomy, and three renames (polars rewrite of #50) #332
  and version: key in model files — and the policy that gives it meaning #67 all wait on The breaking-surface window: sequence #67, #332 and #29 as one decision #432 and nothing said so outside their bodies.

Considered and rejected

A GitHub Project. The account already carries six, one untitled; a seventh
needing manual grooming would drift the same way. The deciding argument is
cadence, not capability: a Project's value is continuous status curation, and
neither a solo maintainer nor an agent invoked on demand supplies that.

track:1–track:4 labels. Tried, and reverted within the hour — they were a
hand-mirror of a markdown table, covered only 20 of 47 open issues, were not a
partition (three issues sat in two tracks), and three of them were simply wrong
because the extraction ran past a section boundary. Sub-issue parentage does the
same job structurally and cannot drift.

A tool that greps issue bodies for dead symbols. It would have caught all
four rows above, and it is still not worth it: bodies legitimately name deleted
things (#332's argument is that #50 cited _sum_piece), upstream paths, and
proposed names that do not exist yet — so it needs a suppression list, a second
thing to groom. Its payoff also lands at triage, which is when the bodies are
being read anyway.

Verified

ruff check clean, 764 passed, mkdocs build --strict clean. Merges cleanly
with #474 (trial-merged locally); the two touch different sections of the file.

[coderabbitai]

Warning

Review limit reached

@FBumann, you've reached your PR review limit, so we couldn't start this review.

Next review available in: 54 minutes

You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository.

How can I continue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews.

How do review limits work?

CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: bc84afa5-ad90-4e98-ba49-e994da4cb965

📥 Commits

Reviewing files that changed from the base of the PR and between a030b1d and e10ee03.

📒 Files selected for processing (1)

• CONTRIBUTING.md

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[codspeed]

Merging this PR will improve performance by 11%

⚡ 1 improved benchmark
✅ 13 untouched benchmarks
⏩ 10 skipped benchmarks¹

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
⚡	test_emit[profiled-s-lpspec-highs]	36.9 MB	33.3 MB	+11%

Tip

Curious why this is faster? Comment @codspeedbot explain why this is faster on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing docs/issue-conventions (e10ee03) with main (603a406)}

Footnotes

1. 10 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩