Skip to content

refactor(adhd,evals): drop one self-check line and add outcome cases - #13541

Merged
kyle-sexton merged 11 commits into
mainfrom
refactor/drop-self-check-lines-batch-2
Nov 5, 2026
Merged

kyle-sexton merged 11 commits into
mainfrom
refactor/drop-self-check-lines-batch-2

Conversation

@kyle-sexton

@kyle-sexton kyle-sexton commented Nov 5, 2026 •

Copy link
Copy Markdown
Member

No related issue: owner decision in the poteto session, 2026-11-05 (no new issues while issues are driven to zero)

Summary

Batch 2 of removing self-check instructions, covering eight plugins: ai-briefing, harness-config, education, ai-slop, adhd, evals, code-tidying and architecture. One cut ships, adhd:shape, with a before/after eval pair that shows no loss. The evals:design cut was tried and reverted, because its eval pair did not show no loss (see Verification), so the line at plugins/evals/skills/design/SKILL.md:215 stays. Three cuts are restored until they can be evaluated on a host where plugin eval can grant Bash. Two new outcome cases are added.

Owner decision (2026-11-05, poteto session), verbatim: "Go w/ recomendations" on this option: "Remove the lines that ask the model to re-read or double-check its own answer, from every skill and agent. Keep the steps that run real tests, builds or type-checks. Put the verifier-subagent steps behind the new verification setting."

Basis: Prompting Claude Opus 5: "Claude Opus 5 verifies its own work without being told to. If your prompt contains explicit verification instructions ("include a final verification step for any non-trivial task," "use a subagent to verify"), remove them: instructions like these cause over-verification on Claude Opus 5, and removing them reduces wasted tokens with no loss in quality."

Repo rule: docs/review/skills-and-agents.md:6-8 says a cut to an existing skill ships only with eval evidence of no loss.

Fix

Line numbers are from origin/main at 2f59a2505c80e8f8c2d8259e50f3bcc3281857b5. None of the eight plugins changed upstream through b489e435c7cc15ea6db127b2d826b970c1039dd7.

Removed (1)

file:line removed text why it is a self-check
plugins/adhd/skills/shape/SKILL.md:205-206 "Then check: reading only the first line and the last line, does the reader know (a) what to do next and (b) what just happened? If yes, send." It re-reads the draft against rules the skill already states: 1 "Lead with the next action" (:52), 3 "End with one concrete next action" (:75) and 5 "Restate state every turn" (:92)

Restored (1)

plugins/evals/skills/design/SKILL.md:215 keeps its main text, "Before finishing, confirm and". The cut was tried and reverted at 8b1c91d89d (see Verification).

Pending eval (host without Docker links)

These cuts are identified as SELF but restored to their main text exactly, because their cases grant Bash and this host cannot run them (see Verification).

file:line self-check text case for the later run
plugins/ai-slop/skills/audit/reference/rewrite-guide.md:261-262 "Then read it once more; what that second read finds is fixed or reported, and the self-audit ends there." (A cut keeps the self-audit's one-pass bound as a drafting rule) audit-fix-self-audit-vocabulary (new, in this PR)
plugins/code-tidying/skills/tidy-diff/SKILL.md:77-78 "Finish with the reference's closing read, bounded as it says." tidy-diff-behavior-changing-cleanup-proposed (existing)
plugins/code-tidying/skills/tidy-diff/reference/generated-code-residue.md:17-19, :28 "Close with one read of the listed lines against every kind below ... then read once more; what the second read finds is reported, and the pass ends there." and ", or that the closing read flags," (A cut keeps "Each kind still present after a proposal is named beside that proposal.") tidy-diff-behavior-changing-cleanup-proposed
plugins/harness-config/skills/audit-instructions/SKILL.md:406-408 "It also confirms the run added zero new interactive gates (report-only contract unchanged; ..." (A cut keeps "the target-model fail-loud stop is an invocation-time validation abort, not an interactive gate, since it prompts nobody and blocks nothing mid-run" as its own sentence) ai-skills-scope-report-only (existing)

For each plugin, build a before copy at 2f59a2505c with the same frozen cases and an after copy with the cut applied, both outside any git repository. Read the case's scaffold.sh, then run both at the same time:

env TMPDIR=<dir of 24 characters or fewer> claude plugin eval <copy>/<plugin> --trust-plugin --keep-temp --judge-model opus --json <repo>/plugins/<plugin>/evals/results/batch2-<before|after>.json --threshold 0.8 --max-cost-usd 5 --no-publish --ablation none --scaffold --allow-tools Bash,Edit,Write --case <case>

Then run run-validity.py <json> --runs 3 on each result, and noise-report.py <after.json> --threshold 0.8 --baseline <before.json> --margin 0.05.

Kept borderline lines

file:line (origin/main) text class
plugins/adhd/skills/clarify/SKILL.md:25 "Done when every decision from the target appears in the restructured view and every omission is named." task specification
plugins/adhd/skills/shape/SKILL.md:197-203 "Before sending, delete: ..." (announcing first sentence, "anything else?" closer, "by the way" sidebar, hedging adverbs) task specification: the hedging-adverb rule is stated only here
plugins/ai-briefing/skills/generate/SKILL.md:224-230 Quality gates: "Every factual item has a source URL and date", and the rest of the list EVIDENCE / task specification
plugins/ai-briefing/skills/generate/reference/build-pipeline.md:117 "Drops partisan-only items at briefing-emit time AND re-validates at slides-emit (defense in depth)" RUN: briefing.md can be edited and re-run between the two emits (:120), and slide-generation.md:160-164 defines its own slide-time filter
plugins/ai-briefing/skills/generate/reference/build-pipeline.md:199 "Drop partisan-only items (already done at briefing-emit time per SKILL.md, but re-validate at slides-emit)" RUN, same reason
plugins/ai-slop/skills/audit/SKILL.md:297-299 "Close each file with the guide's self-audit pass ... before handing it to verification." task specification: the self-audit is the only step that applies the judgment rows to sentences no rule flagged (rewrite-guide.md:133)
plugins/ai-slop/skills/audit/reference/rewrite-guide.md:105-107 "The voice pass and the closing reread below come last" task specification, same reason
plugins/ai-slop/skills/audit/reference/rewrite-guide.md:258-260 "Last pass over each rewritten file ... read it asking 'what still makes this read machine-written?'" task specification, same reason
plugins/ai-slop/skills/audit/context/persist-findings.md:63 "the composed cells must be re-checked" task specification: a cell-escape rule
plugins/architecture/skills/record-decision/SKILL.md:88 "Done when it is indistinguishable in form from the records already there." task specification
plugins/architecture/skills/record-decision/SKILL.md:90 "Done when every heading in the set is one the observed shape uses." task specification
plugins/architecture/skills/improve/actions/deepening.md:66 "Phase 1.5: Verify before publishing" (reproduce each strong candidate against the code) RUN
plugins/architecture/skills/improve/actions/conventions.md:54 "Verify before presenting ... open each cited instance and confirm it" RUN
plugins/code-tidying/skills/batch-simplify/context/reference.md:79 "so the agent self-verifies before returning" RUN: the agent runs the group's verification step
plugins/education/skills/explain/SKILL.md:41 "Done when a rung 1 answer has an analogy whose parts map to the object's ..." task specification
plugins/education/skills/explain/SKILL.md:46 "Done when the reply ends with that single handoff line ..." task specification
plugins/education/skills/explain/SKILL.md:106 Feynman gap check: "ground harder (re-read the source, fetch the primary reference)" EVIDENCE
plugins/education/skills/quiz-me/SKILL.md:87 "Without a fresh sub-agent surface, re-derive the key from the artifact alone before embedding it." INDEP fallback (the verification-setting PR owns these)
plugins/education/skills/teach/context/glossary.md:63 "re-check terms from fast-moving domains per SKILL.md 'Staleness'" EVIDENCE
plugins/evals/skills/plugin-eval/reference/case-authoring.md:3 "Every item is a check to make before a run" task specification: an authoring checklist
plugins/harness-config/skills/audit-automation-gaps/SKILL.md:89 "keep a phase checklist and tick each phase ..." task specification: progress tracking
plugins/harness-config/skills/audit-automation-gaps/SKILL.md:205 "Where the environment exposes no subagent surface, run the gates inline ..." INDEP fallback
plugins/harness-config/skills/audit-automation-gaps/SKILL.md:408 "Where no subagent surface exists, review inline and record that the review was same-context" INDEP fallback
plugins/harness-config/skills/audit-instructions/SKILL.md:375 "Phases B and C require fresh-context, non-fork subagent dispatch. When the Agent tool is blocked ..." INDEP fallback
plugins/harness-config/skills/audit-instructions/reference/conflict-criteria.md:525 "Check first that the mechanism resolves the pair rather than breaking one side." task specification: judging a proposed fix
plugins/harness-config/skills/audit/SKILL.md:377 "Done when every applied fix has a line in the report." task specification

No line with fresh-context dispatch wording or a fresh-eyes-exempt marker was edited.

New eval cases

  • plugins/evals/evals/design-grading-hygiene-recorded: the criteria doc names a grader model other than the one under test and states the re-run cost. Regex graders, with labeled samples. grader-model-differs also accepts a judge named directly ("graded by Claude Opus") and rejects one naming the generator ("graded by Claude Sonnet").
  • plugins/ai-slop/evals/audit-fix-self-audit-vocabulary: a fix removes model vocabulary the detector does not flag ('pivotal', 'delves', 'leverages', 'showcases') and the chat closers. It keeps each fact, and each fact has its own grader (fact-streaming, fact-2gb, fact-dry-run), the way audit-fix-plain-words does it. Regex graders on the file, each with labeled samples.

Fragments: adhd (patch), evals (none), ai-slop (none, for the new case).

Verification

  • plugins/skill-quality/scripts/check-skill.sh over the skills/ root of all eight plugins, on the final tree: 44 passed, 0 failed.
  • validate-cases.py (/evals:validate) on each plugin's evals/, run one plugin at a time: exit 0 and 0 FAIL for all eight, the two new cases included.
  • scripts/run-plugin-tests.sh over the 97 suites in the eight plugins, before the restores: 4 harness-config suites failed (audit-permission-grants/scripts/permission-rule-check.test.sh, audit/scripts/fix-plugin-drift.test.sh, audit-permission-state/scripts/no-writes.test.sh, audit-permission-state/scripts/permission-state-primary-dir.test.sh). This PR changes nothing in their directories, and no-writes.test.sh also fails in the main checkout. Everything else passed. No ai-slop test was rerun.
  • No eval or test asserts the removed adhd line (a repo-wide grep for "first line and the last line" finds only that line).
  • The evals:design cut was reverted at 8b1c91d because no-loss was not shown (7/9 vs 6/9 runs across three passes, one pass 1.00 -> 0.67), per docs/review/skills-and-agents.md:6-8.

Eval evidence

Each pair ran at the same time with the same flags, on plugin copies outside any repository: before = 2f59a2505c with the same frozen cases, after = the branch with that cut applied. Flags: --trust-plugin --keep-temp --judge-model opus --threshold 0.8 --max-cost-usd 5 --no-publish --ablation none, runs 3, --case <case>. The comparison is before against after, so only the with-plugin arm ran.

Skill Case Before After Validity Compare verdict Outcome
adhd:shape sh-multistep-action-first (existing; regex graders on the first and last line) 1.00 (3/3 runs) 1.00 (3/3 runs) VALID both (warning: no Skill-tool call, since the prompt invokes /adhd:shape directly) n too small to call (1 case) Ships. Both versions score 1.00, so the case shows no loss, though it sits at the ceiling and could not show a gain
evals:design, pass 2 design-grading-hygiene-recorded (new) 1.00 (runs 1.0, 1.0, 1.0) 0.67 (runs 1.0, 0.5, 0.5) VALID both n too small to call (1 case) Loss: two of three runs without the line did not state the re-run cost, so the line was restored
evals:design, pass 3 (after the review fix to grader-model-differs) same case 0.83 (runs 0.5, 1.0, 1.0) 1.00 (runs 1.0, 1.0, 1.0) VALID both n too small to call (1 case) No loss in this pass, but across the three passes no-loss was not shown, so the cut was reverted and the line stays

Pass 1 was INVALID (a skill-fired grader cannot pass on a slash-invoked prompt, so it was removed). On its other graders that pass scored 2 of 3 runs passing on both versions. Pass 3's after copy is the before copy with only line 215 changed.

Read these numbers with care. grader-model-differs passed in every run of every pass, so the grader fix does not explain pass 2's loss. Every difference between versions comes from rerun-cost-stated. Across the three passes, that grader passed 7 of 9 runs before the cut and 6 of 9 after it. That is within the run-to-run noise these 3-run passes show: before scored 1.00 in pass 2 and 0.83 in pass 3. One case cannot establish non-inferiority, as each compare verdict says ("n too small to call").

Cost: adhd 0.25 USD per version; evals 0.39 to 0.45 USD per version per pass (list-price estimates).

ai-slop, code-tidying and harness-config were not run. Their cases grant Bash, and plugin eval refuses a Bash-granting run while ~/.docker holds symlinks (contexts and features.json point under /mnt/c/). The owner decided against a link swap, so those cuts wait for a host without the links.

Related

🤖 Generated with Claude Code

kyle-sexton and others added 6 commits November 5, 2026 10:51
Remove the instructions that tell the model to re-read or re-check its
own output with no new evidence: adhd shape's first-line/last-line
check, ai-slop's second read after the self-audit, tidy-diff's closing
read and second read, evals design's confirm-before-finishing, and
audit-instructions' attestation that the run added no interactive
gates. Opus 5 verifies its own work without being told to.

Task specifications, run steps, independent verifiers, gates and
evidence rules stay. ai-briefing, education and architecture had no
line to remove.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Add design-grading-hygiene-recorded, which checks that the criteria
doc names a grader model other than the one under test and states the
re-run cost, and audit-fix-self-audit-vocabulary, which checks that
the self-audit removes vocabulary the detector does not flag.

Keep the rules the removed sentences carried as drafting rules: the
self-audit is one pass, and the target-model fail-loud stop is not an
interactive gate.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The before/after run of design-grading-hygiene-recorded scored 1.00
with the line and 0.67 without it: two of three runs without it did
not state the re-run cost. That is a drop of a third, so the cut does
not ship.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…elf-check cuts

A cut to a shipped skill needs eval evidence of no loss, and these
skills' cases grant Bash, which plugin eval cannot run on this host.
Restore rewrite-guide.md, tidy-diff and generated-code-residue.md, and
audit-instructions to their main text. The new ai-slop case stays for
a later run; its fragment becomes a none fragment.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@kyle-sexton kyle-sexton changed the title refactor: drop self-check lines from adhd, ai-slop, code-tidying and harness-config refactor(adhd): drop shape's first-line and last-line re-check, add outcome cases Nov 5, 2026
@kyle-sexton
kyle-sexton marked this pull request as ready for review November 5, 2026 16:34
@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Nov 5, 2026 •

Copy link
Copy Markdown

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review Status Commit Review trigger
📝 Code Review ✅ Completed 2026-11-05T16:36:38.425179Z 3e191bf Draft marked ready
ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 3e191bf674

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

@github-actions

github-actions Bot commented Nov 5, 2026 •

Copy link
Copy Markdown
Contributor

claude-security-review reviewed 8b1c91d89d5e7f2704bde58be6e8b0df4bc428ab: read 1 of 1 changed file(s); 0 finding(s), 0 as inline comments.

Security review: PR #13541, incremental (fc511be → 8b1c91d)

No security issues found. I read the whole change in the one file in incremental scope. This was a blocking-only review because the lane already reviewed an earlier head, so it would have reported only CRITICAL or IMPORTANT findings. There were none.

What changed: .changes/evals/refactor-drop-self-check-lines-batch-2-650d3d45.md

  • The change fragment's front matter goes from bump: patch to bump: none.
  • The ### Changed release-note text is replaced by a one-line "No release needed (test)" note that describes the new design-grading-hygiene-recorded eval case.

This is release-bookkeeping data. Nothing in it runs as code. It touches no plugin manifest, workflow, trust control, merge hold, token scope, or trusted path. It also does not delete, narrow or soften any standing instruction.

Other checks:

  • Neither the PR context file nor the prior-findings file contains embedded instructions aimed at this lane.
  • The one prior post from this lane, the summary at fc511be, reported 0 findings, so there is nothing to carry forward.
  • The new fragment says "the design skill text is unchanged". The PR description says the evals:design cut was re-applied, and it lists the evals fragment as a patch. Whether that release level is correct is a release-accuracy question for the code-review lane. It is not a security question.

critical-findings: 0

@github-actions github-actions Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

claude-review: 1 inline finding(s) on 3e191bf6743d753257133a944968bd9c3676460b; the summary is in this lane's pull request comment.

claude-lane: claude-review

@github-actions

github-actions Bot commented Nov 5, 2026 •

Copy link
Copy Markdown
Contributor

claude-review reviewed 8b1c91d89d5e7f2704bde58be6e8b0df4bc428ab: read 1 of 1 changed file(s); 0 finding(s), 0 as inline comments.

Code review: PR #13541 (incremental, fc511be3 → 8b1c91d8)

No findings. I read the one hunk of the one file in scope, .changes/evals/refactor-drop-self-check-lines-batch-2-650d3d45.md. This lane already reviewed an earlier head, so I reported blocking findings only.

What changed, and what I checked:

  • The evals fragment changed from bump: patch with a ### Changed entry to bump: none with a "No release needed (test): …" line. The new fragment's claim that "the design skill text is unchanged" matches the code:
    • The full PR diff no longer touches plugins/evals/skills/design/SKILL.md.
    • At head, plugins/evals/skills/design/SKILL.md:215 still reads "Before finishing, confirm and record in the criteria doc:".
    • For evals, the PR now adds only the design-grading-hygiene-recorded eval case. bump: none fits that, and the fragment uses the same "No release needed (test):" form as the ai-slop fragment in this PR.

Note, not counted as a finding: the PR description is now out of date in three places:

  • The "Removed (2)" table lists the evals:design cut at SKILL.md:215 as re-applied and shipping.
  • The Summary says "Two cuts ship, adhd:shape and evals:design".
  • It lists "Fragments: … evals (patch)".

At this head only the adhd:shape cut ships, and the evals fragment is none. The squash-merge message comes from the title and body, so the title ("drop two self-check lines") and these sections should be updated to say the evals:design line was kept.

Not verified by me. The eval scores and validation results in the PR body are the author's claims. I ran nothing.

No plugin manifest changed, so there are no manifest lines. This push changes only a changelog fragment and adds no run spend.

kyle-sexton and others added 2 commits November 5, 2026 12:59
…n's cut

grader-model-differs now accepts a judge named directly, such as
"graded by Claude Opus", and rejects one naming the generator.
audit-fix-self-audit-vocabulary gives each fact its own grader,
fact-streaming among them, with labeled samples for every grader.

The re-run of design-grading-hygiene-recorded scored 0.83 with
design's confirm-before-finishing line and 1.00 without it, so the cut
is re-applied.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@kyle-sexton kyle-sexton changed the title refactor(adhd): drop shape's first-line and last-line re-check, add outcome cases refactor(adhd,evals): drop two self-check lines and add outcome cases Nov 5, 2026
kyle-sexton and others added 3 commits November 5, 2026 14:06
Three eval passes did not show the cut is free of loss: the
re-run-cost grader passed 7 of 9 runs with the line and 6 of 9
without it, and one pass dropped from 1.00 to 0.67. A cut ships only
with eval evidence of no loss, so the line stays and only the new
eval case ships for this plugin.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The fragment front matter takes only bump; the reason moves to the body,
matching the ai-slop fragment in this change.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@kyle-sexton

Copy link
Copy Markdown
Member Author

Update: the evals:design cut is reverted at 8b1c91d. Across three passes the re-run-cost grader passed 7/9 with the line and 6/9 without it, and one pass fell from 1.00 to 0.67, so no-loss is not shown and the line stays (docs/review/skills-and-agents.md:6-8). This PR now ships the adhd:shape cut, the grader fixes and the two new eval cases; the evals fragment is bump: none.

@github-actions

github-actions Bot commented Nov 5, 2026

Copy link
Copy Markdown
Contributor

The pr-verify verifier failed head 8b1c91d89d5e7f2704bde58be6e8b0df4bc428ab. What it wrote, as data:

What it wrote (25 lines)
Summary: There is no linked issue, so the spec is this pull request's own title and body (#13541).
  The diff does not match it. The title and body say two self-check lines are removed: adhd:shape
  and `plugins/evals/skills/design/SKILL.md:215`. They also say the evals fragment is a patch bump.
  The diff removes only the adhd:shape line (`plugins/adhd/skills/shape/SKILL.md:205-206`). It
  leaves the evals:design skill unchanged, and its evals fragment is `bump: none`. The author's
  later comment says the evals:design cut was reverted at the judged head because no-loss was not
  shown, which the repo rule requires (docs/review/skills-and-agents.md:6-8). Comments are context,
  not spec, and the title and body were never updated, so the delivered change and its description
  disagree. The rest checks out. The adhd removal is coherent: rules 1, 3 and 5 still carry the
  first-line and last-line requirements. The two new eval cases (ai-slop
  audit-fix-self-audit-vocabulary and evals design-grading-hygiene-recorded) have regex graders
  whose labeled pass and fail samples match as intended, and the scaffold only writes a fixture
  file. All CI jobs that cover the changed files concluded success (test-bash, test-python,
  check-skills, check-plugins, lint-shell;
  https://github.com/melodic-software/claude-code-plugins/actions/runs/42091214386). test-node was
  skipped, which is expected because no Node files changed. The capture was blocked (claude was not
  on PATH), so it is no evidence either way. There are no binary files, no secrets, no takeover-set
  files and no loosened holds.
- PR #13541 title and body: they still describe the reverted evals:design cut. Change the title from
  'drop two self-check lines' to one line. In 'Removed (2)', drop the
  `plugins/evals/skills/design/SKILL.md:215` row, or move it to the restored/pending lines. Change
  'Fragments: ... evals (patch)' to evals (`none`), and change 'Two cuts ship' and the pass-3 'cut
  is re-applied' outcome so they say the line stays. Without these changes the spec and the diff
  disagree. The alternative is to re-apply the cut in the diff, but by the author's own eval
  evidence that would break the repo's no-loss rule.

This pull request leaves the merge queue, and ci-status holds it out while it carries pr-verify-failed. A pr-verify pass removes the label, at a new commit or at this same head: once a trusted person edits the body, and only trusted people have edited it since the fail, this head is judged again, so a fail on the body or the issues it links needs only the body fixed, no new commit.

@github-actions github-actions Bot added the pr-verify-failed pr-verify failed this head; a pass at a new commit, or after a trusted body edit, removes it. label Nov 5, 2026
@kyle-sexton kyle-sexton changed the title refactor(adhd,evals): drop two self-check lines and add outcome cases refactor(adhd,evals): drop one self-check line and add outcome cases Nov 5, 2026
@github-actions github-actions Bot removed the pr-verify-failed pr-verify failed this head; a pass at a new commit, or after a trusted body edit, removes it. label Nov 5, 2026
@kyle-sexton
kyle-sexton added this pull request to the merge queue Nov 5, 2026
Merged via the queue into main with commit ef33578 Nov 5, 2026
54 checks passed
@kyle-sexton
kyle-sexton deleted the refactor/drop-self-check-lines-batch-2 branch November 5, 2026 20:16
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant