max / makenotwork

Fix the clippy and test failures the newly configured gates found None of these were regressions; they were caught the first time the gates that cover them actually ran on the Sando host (rebuild 27). clippy, five errors in makenotwork. Three redundant references in format! arguments, in bulk item copy and the two API-key masking sites; two assert! equality comparisons in tier_prices that should be assert_eq!, which also gives a useful message on failure instead of just "assertion failed". The two test failures were the same bug twice, and it belongs to c8ff4d2b ("Spell the product name Makenotwork everywhere it is displayed"). That commit renamed the displayed string and missed two assertions that were still matching the old one: - admin::health_gates_full_dashboard_to_admins asserted the anonymous status page contains "MakeNotWork". The page correctly renders "Makenotwork". The brand doc lists "MakeNotWork" as a spelling never to use, so the test was asserting the forbidden form. - exports::content_export_zips_files_and_uploads_to_s3 asserted the export README manifest starts "Makenot.work Content Export"; it is "Makenotwork Content Export" (src/routes/api/exports/content.rs:416). Fixed the assertions rather than the code: in both cases the code was right and only the expectation was stale. Two forbidden spellings survive on purpose. tests/workflows/sso.rs's module comment is prose about the delegated-login button, and src/auth.rs:887 sends "MakeNotWork-Security-Check" as an outbound User-Agent, which is an identifier a remote service may match on rather than displayed copy. Neither is UI. Green under the gate's own invocations: cargo clippy --all-targets --features fast-tests -D warnings, cargo fmt --check, and the full suite (1912 unit, 1215 integration, 0 failed).