[aliromman]

The shared research workflow discovers test_research_*.py. Four test modules from #5726, #5731, #5732 and #5733 used other names, so that command omitted their 22 tests. Rename the four files to match the existing pattern. All file contents remain byte-identical.

The original contributions were integrated while their final checks were in progress. This follow-up starts from canonical main c297a40e7732bc9452bbf745f9aad43455b1e330; it completes their CI registration. The 15 original contributions remain independently reviewable in Discussion #5722.

Validation: all 69 new tests from those contributions pass together on current main with these renames, under normal and optimized Python 3.12. Git reports four 100% renames, with zero inserted or deleted lines. git diff --cached --check passes.

The final upstream refresh read revision 8 at de5ec2e7a6074068452f45b1a191b5c72832df4a. Its verification-tooling priority still fits this correction, and the selected record is unchanged. The live overlap check covered all 309 open PR heads and bodies and the 10 most recently updated Discussions. No matching scope was found. An additional file-list refresh stopped at historical PR #3120 with a GitHub HTTP 500; the metadata check is not a proof of novelty.

Prepared for aliromman with OpenAI Codex assistance. Preserve the original tool, test and predecessor credits. This changes test discovery only and makes no mathematical or selected-frontier claim.

Verified head: e3117af2cd7d1e5e473ee1ff73c94ebbe8daa16f. Hosted results are recorded below.

The broader local discovery run completed 526 tests in 432.374 seconds with 5 failures, 16 errors and 29 skips. This is not a full local pass. Diagnostics include unavailable historical Git blobs, omitted sparse-checkout inputs, a denied ps call, a 30-second inventory CLI timeout, and a current-assignment comparison failure in the sparse checkout. The four renamed modules and all 69 new contribution tests passed separately. The hosted full-checkout result is recorded below.

Hosted verification at head e3117af2cd7d1e5e473ee1ff73c94ebbe8daa16f: the plan log records all 69 new contribution tests as passing. The full run completed 533 tests with 8 failures, 2 errors and 22 skips. Every failure/error is in the existing test_research_current_assignment.py after upstream changed the catalogue to revision 8. The old tests still require 16 rewritten cards and legacy receipt fields; the revision-8 cards already use the current entry command. Two failures (0 != 16 and the missing-helper expectation) reproduce from exact unchanged main de5ec2e7a6074068452f45b1a191b5c72832df4a, without this rename patch. The plan is red and the replay job is skipped; the new tests passing does not make the full plan pass.

The isolated inventory-coverage recheck passes all 5 tests (41.958 seconds); this resolves the transient 30-second timeout from the broad local run. Other broad-run sparse-checkout and sandbox limitations remain as recorded.

Remaining handoff: update the current-assignment tests for the live revision-8 catalogue while retaining explicit frozen legacy-routing fixtures, then rerun the shared plan. This is outside the four-file rename patch and does not require a frontier change.