diff --git a/.verify-floor.json b/.verify-floor.json
index c59670a..7934b95 100644
--- a/.verify-floor.json
+++ b/.verify-floor.json
@@ -2,8 +2,9 @@
  "collected": 502,
  "passed": 502,
  "skipped_max": 26,
+ "mirror_skipped_max": 31,
  "selftest_skipped_max": 7,
  "gate_skipped_max": 2,
- "note": "Recorded by verify.py --update-floor, except the *_max ceilings, which are edited BY HAND and never re-measured. `collected` catches tests that stopped being collected; `passed` is compared against passed+skipped, so a check may move between passing and consciously-skipped but the two together may never shrink. The *_max ceilings bound the skipped side: 24/7/2 is exactly what a machine with none of this instance's local prerequisites skips (a GitHub runner: no agent CLIs, no ~/.codex/skills, no /Applications/ChatGPT.app, no populated capability ledger), measured 2026-08-21. On the owner's machine all prerequisites exist and nothing skips at all. Raising a ceiling is a deliberate act: it means agreeing that one more thing is allowed to go unchecked, so say which and why in the commit. LOWERED 26 -> 24 on 2026-08-22, reverting the raise made earlier the same day. The two kill-switch exemption tests no longer need to skip on a bare runner: their declarations moved out of the running instance's ledger and into capabilities.KNOWN_DECLARATIONS, so they assert code-derived truth and run everywhere. Moving a test back below the ceiling is the preferred way to lower it -- fix what made it machine-dependent, rather than agreeing to check less. FLOOR 345 -> 353 on 2026-08-22: 345 was measured on a branch cut before #13 (research panels/rounds/domain studies) merged, so the recorded floor sat 8 tests BELOW what main actually collects. A floor below reality is the permissive direction -- those 8 could have silently stopped being collected and still cleared the check, which is exactly the hole this file exists to close. Measure the floor on the merge result, not on the branch. Raised again on 2026-08-22 by the producer-identity-scope branch, which adds tests on top of the 353 recorded by #15; re-measured after rebasing rather than assumed. NOTE: `verify.py --update-floor` REPLACES this note with a generic one, so it must be restored by hand after every use \u2014 the ceiling rationale is the only record of which prerequisite justifies each skip. FLOOR 365 -> 366 on 2026-08-22 (heartbeat-ordering work, PR #18): exactly one new test, test_capabilities.test_no_tick_producer_runs_above_the_heartbeat_export. No ceiling moved and nothing new is skipped -- it reads source files rather than a populated ledger, so it runs on any machine. The branch recorded 354 because it was cut before #16 merged; re-measured on the MERGE RESULT per the rule above, which is exactly the mistake that put the floor 8 below reality last time. FLOOR 366 -> 368 on 2026-08-23: main collected 368 while this file recorded 366, drift left by #34 (evidence-acquisition landed, +1) and #37 (tick capability evidence, +1) whose authors each measured against a branch cut before the other merged. A floor BELOW reality is the permissive direction this file exists to close -- those two could have silently stopped being collected and still cleared the check. Measured on the merge result per the rule above: 368 passed, 0 failed, 0 skipped, 83/83 selftests, 43/43 can-fire, 5/5 gates. CEILING 24 -> 26 and FLOOR 368 -> 387 on 2026-08-23 (profiles/provenance branch, PR #42). This file CONFLICTED with #50, which raised the floor 366 -> 368 on main while this branch raised it to 387; resolved as the UNION rather than by taking a side -- #50's rationale is retained above and the count was RE-MEASURED on the new merge result instead of keeping either number. 368 (main) + 19 (this branch's net new tests) = 387; #50 corrected recorded drift rather than adding coverage, which is why 387 is unchanged from the pre-conflict measurement. Measured in a runner sandbox reproducing CI exactly (361 passed, 26 skipped, 387 collected) AND on the owner's machine (387 passed, 0 skipped, 5/5 gates). The two new skips are drift detectors against a REAL installed agent runtime, so neither can be moved below the ceiling -- the preferred way to lower one: (1) agy advertised-models cache absent, since comparing declared model ids against the catalogue agy actually advertises needs that catalogue, and a fixture would exercise the comparison while detecting no real drift; (2) vibe config absent (~/.vibe/config.toml), since active_model cannot be read to check for drift when there is no config to read. Both name their missing prerequisite, so a green run still states what it did not check. A third candidate skip was REFUSED: dispatcher's per-run agy-log assertion failed on a bare runner because adapters.advertised_models shells out to `agy models` when its disk cache is cold, and that probe landed inside a monkeypatched subprocess.run and overwrote the captured command. That is a stub leak, so it was fixed by ISOLATING the double rather than by skipping -- which makes CI run MORE. FLOOR 387 -> 391 on 2026-08-23 (improvement-log accessor, PR #59): exactly four new tests, all in test_improvement_log.py -- three read tracked files in the tree (the pointer's size and content, and that CLAUDE.md 0 step 3 and 5 name the accessor rather than a bare path) and one runs the accessor as a subprocess against a path that cannot exist. None reads a populated ledger, an agent CLI or ~/.codex, so all four RUN on a bare runner and NO ceiling moved: nothing new is skipped. Measured on the MERGE RESULT after rebasing onto origin/main af6654d, which collected 387 -- not on the branch base, per the rule above. FLOOR 391 -> 402 on 2026-08-23 (Gate python-ci configuration, the PR that adds the missing .github/workflows/autofix-versions.env): exactly 11 new tests, all in test_ci_gate_config.py, which read committed files only -- the pin file, ruff.toml, mypy.ini, pr-00-gate.yml's toggle annotations and docs/CI_LINT_BASELINE.md. NO ceiling moved. On any CHECKOUT -- CI, the owner's tree, a second instance -- all 11 run: they need no installed linter and no populated ledger. In the EXEC-MIRROR layout all 11 skip with one named reason, because orch-sync-mirror.sh copies root-level *.py only, so .github/workflows, docs/ and scripts/ are genuinely absent there (env_prereq.repo_files_absent). That lands at 11/26 on a machine that otherwise skips nothing, and CI stays at 26/26, so no ceiling needed raising. The skip gate is the presence of those DIRECTORIES, never of the pin file itself -- gating on the file would have made the test that checks for it unable to fail. Measured on the merge result, twice: the branch was rebuilt on origin/main after #42 and #59 merged, and re-measured after #61 merged and was merged in -- 393 passed + 9 skipped = 402 collected both times, so #61 added no collected tests and this floor is not sitting below reality. #61 itself left main's floor at 391, which is exactly main-without-these-11, so there is no inherited drift to correct. RULE CHANGE 2026-08-23: `collected` is now an EQUALITY, not a minimum. Every floor entry above this one records the number being found BELOW reality and hand-raised after the fact -- 21 low at the worst, then 8, then 1, then 2 -- because nothing ever required a test-adding PR to touch this file, so the permissive direction was silent by construction and the rule 'measure on the merge result' had to be restated three times with nothing enforcing it. verify.py now FAILS when collected exceeds the floor, printing the two integers to write. That also makes the concurrency case self-enforcing: once every test-adding branch must edit these same two lines, two concurrent branches CONFLICT IN GIT, so the second cannot merge without rebasing onto the first and re-measuring on the actual merge result. Demonstrated repeatedly on the change itself: six merges landed on main in the two hours it took to write, moving this file 368 -> 387 -> 391 -> 402, and every one would have left the floor below reality under the old one-directional rule. `passed` deliberately stays a MINIMUM on passed+skipped: only collection is machine-invariant (a skipped test is still collected), measured across machines at 391 collected on both, with pass/skip splits of 365/26 on CI against 391/0 locally. The *_max ceilings are untouched by this change and nothing new is skipped. `--update-floor` also stops REPLACING this note -- it appends -- so the warning above about restoring it by hand no longer applies; and drift does NOT block --update-floor, since a gate that forbade its own only remedy would be a deadlock (the first draft was exactly that). FLOOR 402 -> 407 on 2026-08-23 (findability admission requirement). (findability admission requirement). (findability admission requirement). (findability admission requirement). Exactly five new pytest tests, all in test_capability_admission.py: test_findability_distinguishes_its_three_sub_causes, test_findability_blocks_new_capabilities_and_reports_older_ones_as_debt, test_unreadable_reach_is_not_evaluated_and_never_a_failure, test_findability_exemption_is_declared_in_code_not_in_a_live_ledger, test_consult_sites_are_falsifiable_claims_about_real_callers. NO CEILING MOVED and nothing new skips: all five build synthetic ledgers in a tempdir or read committed tables, so none needs a populated capability ledger, an agent CLI or ~/.claude/skills. The one machine-dependent thing they touch -- an external consult site declared in capability_advisor.CONSULT_SITES whose skill prompt is not on this machine -- is reported as UNVERIFIED rather than skipped, because absence of the caller is not refutation of the claim; the in-tree site (tick) is asserted verified on every machine so the check can never degrade into 'everything unverified, nothing tested'. Measured on the merge result per the rule above: this file CONFLICTED three times while the branch was open, as main went 387 -> 391 -> 402 (#61, #64, #65, #60). Each time it was resolved as the UNION rather than by taking a side, and the count was RE-MEASURED on the new merge result rather than either number being carried forward: 402 (main at bd6da2e) + 5 (this branch's new tests) = 407. That is the rule this file already states -- measure the floor on the merge result, not on the branch -- and it mattered here, because #60 both deleted test_ci_gate_config.py and added more than it removed, so guessing in either direction would have been wrong. -> re-measured on 2026-08-23 (PR #62, the four deferred #42 review findings): three new tests, all machine-independent (each builds its own tmp_path Brain and manifests), so NO CEILING MOVED and nothing new is skipped. Fourth conflict for this branch, and the first one under the EQUALITY -- which is the point: the equality's own rationale says git conflict detection is what enforces 'measure on the merge result', and that is exactly what happened here. Under the old minimum the three earlier conflicts could each have been resolved by keeping the larger number; under the equality the count MUST be measured, and it was. RESOLVED AGAINST #68 (findability admission requirement) on 2026-08-23: taken as the UNION per the rule this file states -- #68's five-test entry is retained above and this branch's three-test entry beside it -- and the count RE-MEASURED on the merge result rather than keeping either side's number. main fc1fd42 collects 407; this branch adds 3; 410 measured with `pytest --collect-only -q` on the merge result, not assumed. Ceilings untouched at 26/7/2 and nothing new is skipped. Also resolved in the same merge: langsmith-fleet-worker-attempt.json, a CI-emitted `langsmith-fleet/v1` worker-attempt record whose two sides differed only in `emitted_at` and `pr_number` (62 here, 68 on main). Main's NEWER record was kept rather than this branch's older one -- discarding a newer provenance observation to win a merge would corrupt exactly the causal-provenance evidence CLAUDE.md 2 protects, and this branch's own run re-emits its record anyway. FLOOR 410 -> 411 on 2026-08-23 (CodeRabbit follow-up on PR #42, thread 3837879039; re-measured again after #56 made `collected` an EQUALITY, which makes an assumed number a hard RED rather than a quiet pass -- main stayed at 402 across #56, and the merge result measures 403, so #56 added no collected tests and this is main's 402 plus this branch's one): exactly one new test, test_feedback_model_provenance.test_late_sweep_completes_terminal_attempts_never_one_in_flight, which pins that ledger_reconcile.resolve_unresolved_worker_attempts completes only TERMINAL unresolved worker attempts and never one still in flight. No ceiling moved and nothing new is skipped -- the test builds its own tmp ledger and codex rollout fixture and monkeypatches adapters.CODEX_SESSIONS, so it needs no agent CLI and no populated capability ledger and runs on a bare runner. RESOLVED AGAINST #59 (improvement-log accessor), which raised the floor 387 -> 391 on main while this branch raised it to 388: taken as the UNION -- #59's rationale is retained above and the count was RE-MEASURED on the new merge result rather than keeping either number, which is the rule this file states and the mistake that once put the floor 8 below reality. 391 (main, incl. #59's four tests) + 1 (this branch's one new test) = 392 measured, not assumed: 392 passed, 0 failed, 0 skipped, 83/83 selftests, 43/43 can-fire, 5/5 gates. Three sibling follow-up branches are in flight against this same main (CI/ruff config, arm-attribution + durability, adapters label->ID); if this file conflicts with one of them, resolve as the UNION and RE-MEASURE on the new merge result rather than taking either number -- that is what #42 and #50 did, and taking a side is what put the floor 8 below reality earlier. RESOLVED AGAINST #68 (findability admission requirement) on 2026-08-23: taken as the UNION per the rule this file states -- #68's five-test entry is retained above and this branch's one-test entry beside it -- and the count RE-MEASURED on the merge result. main fc1fd42 collects 407; this branch adds 1; 408 measured with `pytest --collect-only -q` on the merge result, not assumed. Ceilings untouched at 26/7/2 and nothing new is skipped -- the one new test builds its own tmp ledger and codex rollout fixture, so it runs on a bare runner. Also resolved in the same merge: langsmith-fleet-worker-attempt.json, a CI-emitted `langsmith-fleet/v1` worker-attempt record differing only in `emitted_at` and `pr_number`; main's NEWER record was kept, since discarding a newer provenance observation to win a merge would corrupt the causal-provenance evidence CLAUDE.md 2 protects. FLOOR 411 -> 415 on 2026-08-23 (PR #70 diagnostics salvage): four new collected tests in test_capability_set_coverage.py from the PR #43 salvage plus CodeRabbit follow-ups on PR #51/#70 \u2014 union/missing-candidate fetch command, truncation after six modules, AST-scoped gate-call audit, and entrypoint-diagnosis coverage. NO CEILING MOVED and nothing new is skipped; all inject synthetic ledgers or read committed source. Measured on the merge result at 91d37fa: 389 passed + 26 skipped = 415 collected on CI, not assumed. FLOOR 415 -> 416 on 2026-08-23 (the dangling-citation follow-up, PR #74): exactly ONE new test, test_ci_gate_config.test_every_cited_repo_path_resolves, which reads the two committed config files this repo OWNS (the pin file and ruff.toml) and asserts every repo-relative path they cite exists. It exists because the pin file shipped citing docs/ci/LINT_BASELINE.md when the real path was docs/CI_LINT_BASELINE.md: the sibling checks read that file's CONTENTS thoroughly and its PROSE not at all, and the prose is the only pointer telling a reader where to re-measure before bumping a pin. Scoped to the two owned files deliberately -- scanning pr-00-gate.yml yields six findings that are all correct as written (guarded by hashFiles or a .agents check, or upstream paths), and a test that cries wolf gets waived. NO ceiling moved. RE-MEASURED SIX TIMES as the base moved under this ONE-LINE change: bd6da2e 402 -> ddb0928 402 -> fc1fd42 407 -> 0d661e3 407 -> 0593eeb 411 -> 6fed4ad 415, each +1 with this test, and the branch was rebuilt on each rather than the number carried forward. THIS BRANCH IS THE WORKED EXAMPLE of the equality's concurrency cost, so record it rather than rediscover it: main moved EIGHT times in the ~2.5 hours a one-line comment fix was open (#56, #68, #73, #69, #62, #70 and two direct commits), the floor line conflicted THREE separate times, and two merges overlapped the change directly -- #73 landed a byte-identical copy of the backplane-conformance.yml guard this branch also carried (dropped as redundant), and #69 edited this very test file in a neighbouring region. The equality is still the right call and should stay: every entry above this one records the floor being found BELOW reality, which is the permissive direction. But no amount of author care wins this race, because the correct value is only knowable on the merge result. The durable fix is CI running `verify.py --update-floor` on the merge commit, which keeps the equality and removes the race; until then a test-adding PR must be merged promptly after going green, because it re-conflicts on roughly every subsequent merge. FLOOR 416 -> 427 on 2026-08-23 (PR #72 hygiene untrack, rebased after #71 merged): exactly 11 new tests from test_repo_artifact_hygiene.py with root-anchored gitignore patterns. NO ceiling moved. Measured on merge result after #71 landed on main: 416 (main) + 11 = 427 collected via pytest --collect-only -q, not assumed. #71's simpler untrack landed first; this branch carries the full hygiene test suite and corrected root-anchored patterns. FLOOR 427 -> 428 on 2026-08-23 (PR salvaging #34/#42 remnants): exactly one new test, test_feedback_model_provenance.test_gemini_provenance_reads_the_per_run_log_before_the_conversation_store, recovered from #42's post-merge commit 4e0d6ae along with the adapters catalog work it exercises. No ceiling moved and nothing new is skipped -- it seeds adapters._ADVERTISED_MEMO instead of letting the catalog probe shell out, so it runs on any machine and adds no prerequisite. `passed` is 428 rather than the 426 verify.py suggested on this machine: two test_capabilities liveness tests (test_gate_blocks_execution_is_opt_in_and_narrow, test_evidence_gate_kind_is_not_blanket_observer) currently fail HERE on pristine main as well, because the hourly fleet tick mutated the machine-local ledger and range-lane-rollout now classifies matched_not_invoked rather than deliberately_gated. That is ledger STATE, not this branch and not the code -- CI bootstraps an empty ledger and counts 428/428. Recording 426 would have baked a local environment failure into the floor as though it were the expected result. FLOOR 428 -> 441 on 2026-08-23 (coverage measures what actually runs): exactly 12 new tests, all in test_verify_coverage_mode.py. They read committed files and verify.py's own source, and monkeypatch verify.COVERAGE in-process -- no populated ledger, no agent CLI, no ~/.codex, and no coverage RUN -- so all 12 execute on any machine and NO ceiling moved. The change itself is a measurement fix, not a gate: `verify.py --coverage` wraps each child in `coverage run --parallel-mode` and combines, because the per-module --selftest is a SUBPROCESS and a pytest-only coverage run cannot see it. That blind spot was most of the codebase -- 78 modules have no test_*.py at all, ~85,500 lines, 79.6% of non-test root Python -- so the reported 48.45% was measuring the gap in the instrument, not a gap in the tests. Combined: 76.1% (45,049 statements, 10,774 missed). Twelve of the twelve modules the old report named as worst were selftest-only; outcomes.py reported 9.0% and measures 61.9%, watch.py 9.6% -> 86.9%. Coverage is OFF by default and deliberately never touches the exit code -- one of the 12 tests pins that, because enforcing a threshold here would reward pytest wrappers around already-tested modules: metric up, assurance flat. Measured on the MERGE RESULT: branched from origin/main bcc68cd (floor 427), then REBASED onto f5f1c39 when it landed underneath and re-measured on the new merge result rather than carrying the old number: 428 + 13 = 441. +1 on 2026-08-23 (440 -> 441 after the rebase), same branch: test_the_cli_help_actually_renders. It exists because this branch BROKE `verify.py --help` and its own twelve tests did not notice. argparse interpolates help strings with `% params`, so the literal `~80%` in the --coverage help was read as an `%o` octal conversion and --help died with 'badly formed help string'. All twelve original tests passed: every one inspected source text or monkeypatched a flag, and not one RENDERED the help -- a construction-time test suite that never exercised the constructed thing, which is this repo's founding defect one layer up. CI's verify.py gate caught it, which is the check of last resort working as intended. The new test runs `verify.py --help` as a subprocess and asserts it exits 0, so it runs anywhere and NO ceiling moved. Rendering rather than grepping for `%` is deliberate: a grep would flag the legitimate `%(default)s`. FLOOR 441 -> 442 on 2026-08-23 (matched_not_invoked yields to observers and declared gates): exactly one new test, test_capabilities.test_matched_not_invoked_yields_to_observers_and_declared_gates. It is SYNTHETIC on purpose and that is the point of it: the two tests that caught this bug in the wild read the LIVE ledger, so they skipped with a named reason on the empty ledger ci.yml bootstraps -- the defect was red on every populated machine and green on CI for as long as it existed. A synthetic row asks the same question everywhere, so this one RUNS on a bare runner and NO ceiling moved. The fix itself moves matched_not_invoked below `observing` and below the DECLARED deliberately_gated check in classify_liveness: it was the first check, which made it the fourth instance of the unescapable label the comments in that function exist to fix. Audited before committing -- 12 of 43 live rows reclassify (ten observers to observing, two declared gates to deliberately_gated) and ZERO move for any other reason, so nothing is reclassified by inference and the weaker gate_reason-only branch is untouched. Measured on origin/main 0d9c3a7, whose recorded floor is 441, so 441 + 1 = 442. FLOOR RECORDED by verify.py --update-floor on 2026-08-23: collected=442, passed=442. Ceilings preserved, never re-measured \u2014 they are edited by hand. FLOOR RECORDED by verify.py --update-floor on 2026-08-24: collected=442, passed=442. Ceilings preserved, never re-measured \u2014 they are edited by hand. FLOOR RECORDED by verify.py --update-floor on 2026-08-24: collected=442, passed=442. Ceilings preserved, never re-measured \u2014 they are edited by hand. MYPY_EXEMPT_MAX introduced at 64 on 2026-08-23, the same change that flipped `typecheck` ON in pr-00-gate.yml. It was OFF because 608 whole-tree errors were drainable 0 per PR -- a gate whose clear path is blocked by the thing it measures. Two real changes opened it: the src/ move scoped the Gate's `target=src` to the 99 modules (608 -> 467), and pyproject.toml's [[tool.mypy.overrides]] exempts the modules that still have findings BY NAME so the 35 already-clean ones are checked today. 12 var-annotated findings were then drained (467 -> 455, 66 -> 64 modules) to prove the drain works rather than promise it. This ceiling may only ever be LOWERED, by typing a module and deleting its line -- raising it means agreeing one more module goes unchecked, so say which and why. NOT a suppression: no error code is disabled anywhere and the 455 stay visible via `python3 scripts/ci_lint_baseline.py`. Edited BY HAND like the other *_max values; --update-floor never re-measures it. MYPY_EXEMPT_MAX LOWERED 64 -> 43 on 2026-08-24, the ratchet's first real drain: 21 modules typed clean and removed from the exempt list, 455 -> 430 findings. Targeted the modules with 1-3 errors on purpose -- fixing 50 errors spread across the big modules would move this number by ZERO, and this number is what the gate reads. Lowering it is the drain; it may never be raised without naming the module and why. MYPY_EXEMPT_MAX LOWERED 43 -> 33 on 2026-08-24 (batch 2 of the drain): the whole <=5-finding tail typed clean. 66 of 99 modules now checked. What remains is five per-module campaigns (capability_advisor, dispatcher, capability_propensity, runtime_ac_gate, capabilities hold most of it), so future batches take ONE big module at a time rather than skimming. (batch 2 detail: 43 -> 26, 430 -> 374 findings, 73 of 99 modules checked. `mypy_path` gained `tests` so mypy can RESOLVE the recurrence-fixture roster capability_admission legitimately imports; the target stays `src`, but mypy then follows into that one test file, whose two findings were fixed rather than configured around.) MYPY_EXEMPT_MAX LOWERED 26 -> 20 on 2026-08-24 (batch 3): 79 of 99 modules checked. FLOOR 442 -> 448 on 2026-08-24 (absent-check detector + its ratchet): SIX new tests in tests/test_checks_reported.py holding the frequency rule and the expected-check ratchet. NO ceiling moved -- they call pure functions and read one committed JSON file, so they run on any machine. The ratchet exists because dogfooding the detector caught it DISARMING ITSELF: while pr-00-gate.yml sat held, every merged PR merged without the Gate, so after twelve such merges the Gate's checks fell below the 75% frequency threshold, the expected set eroded 23 -> 14, and PR #91 was pronounced clean by the tool written to catch exactly that. config/expected-checks.json is the high-water mark, seeded from PRs #87/#89 whose Gate demonstrably ran, and it only comes down when somebody deletes a line. LOWERED to 16 on 2026-08-24 (batch 4, re-applied after merging main, which carried #94/#100's own floor work \u2014 main's note kept, only the bound re-set). LOWERED 16 -> 16 on 2026-08-24 (batch 5): 83 of 99 checked. LOWERED 16 -> 12 on 2026-08-24 (batch 5): 87 of 99 checked. Lesson recorded three times now: measure with the PROJECT run, never per-file \u2014 `mypy src/X.py` reports clean for modules the project run still flags. FLOOR 448 -> 450 on 2026-08-24 (giving the per-node deliberate-break finding its consumers): exactly two new pytest tests, both in test_synthesis_promotion.py -- test_a_passing_break_names_the_tautologies_it_carried and test_the_break_caveat_is_silent_on_clean_and_on_pre_per_node_evidence. NO CEILING MOVED and nothing new is skipped: the first builds its own git repo in tmp_path and runs the real verification path (git + pytest only, no populated ledger, no agent CLI, no ~/.codex), and the second is a pure unit on _break_caveat. The third piece of this change is a runtime_ac selftest case, which adds NO collected test -- runtime_ac has a --selftest and synthesis_promotion does not, which is the whole reason only one of the two needed pytest tests. Measured on the merge result: branched from origin/main 8549f84, whose recorded floor is 448, and re-checked that main had not moved before writing this. 448 + 2 = 450, from verify.py's own count (450 passed, 0 failed, 0 skipped, 85/85 selftests, 5/5 gates), not assumed. FLOOR 448 -> 451 on 2026-08-24 (offload guidance correction, PR #113): THREE new tests in test_repo_artifact_hygiene.py -- two parametrized entries adding src/UNKNOWN.egg-info/PKG-INFO and UNKNOWN.egg-info/PKG-INFO to EMITTED_ARTIFACTS, and test_no_build_metadata_is_tracked. NO ceiling moved and nothing new skips: all three shell out to git against this checkout, so they run anywhere git does. They exist because THIS PR carried four committed build artifacts and no check objected. A CI step runs `pip install -e .`, setuptools writes src/UNKNOWN.egg-info/ (named UNKNOWN because pyproject.toml declares no [project] on purpose), and the repo's own autofix bot committed all four -- .gitignore had no *.egg-info/ entry and this suite only knew the langsmith-fleet family. CodeRabbit caught it; nothing local did. Both locations are listed because a checkout builds into src/ while the EXEC MIRROR IS FLAT. The tracked test is separate from the ignored test on purpose: the four files were committed BEFORE the pattern existed, and adding a pattern does nothing to a path git already tracks -- so an ignore-only check passes on a repo still carrying the debris. RESOLVED AGAINST #110 and #114 on 2026-08-24, which moved main 448 -> 450 while this branch raised it to 451: taken as the UNION per the rule this file states -- main's entries are retained above and this branch's three-test entry beside them -- and the count RE-MEASURED on the merge result rather than either number being carried forward. 450 (main) + 3 (this branch) = 453, measured with `pytest --collect-only -q`, not assumed. Ceilings untouched and nothing new is skipped. FLOOR 453 -> 458 on 2026-08-24 (untracking .coverage, the coverage database this repo's own verify.py writes): exactly FIVE new tests, all in test_repo_artifact_hygiene.py -- two parametrized ignore cases over COVERAGE_DATA_FILES (.coverage and a parallel-mode .coverage.<host>.<pid>.<random>), one untracked case, and two parametrized must-stay-committable cases (.coveragerc, tools/coverage_guard.py). NO CEILING MOVED and nothing new is skipped: all five ask GIT about a path in this checkout (check-ignore / ls-files) and need no populated ledger, no agent CLI and no ~/.codex, so all five RUN on a bare runner. The change is `git rm --cached .coverage` plus root-anchored `/.coverage` and `/.coverage.*`; the 90 KB SQLite database arrived on main in #109, a typing PR whose every other file is about mypy, and verify.py's coverage_reset() unlinks it and coverage_combine_and_report() rewrites it on every --coverage run -- so while tracked it was an opaque binary rewritten by the command that produces this repo's verdict. Both patterns because two different steps write them, and `/.coverage.*` rather than `/.coverage*` because the second swallows .coveragerc. Break -> revert performed in all three directions and recorded in the test file. Measured on the MERGE RESULT: fast-forwarded onto origin/main 5c769e0, whose recorded floor is 453, and re-fetched to confirm main had not moved again before measuring. 453 + 5 = 458 from verify.py's own count (458 passed, 0 failed, 0 skipped, 85/85 selftests, 5/5 gates), not assumed. LOWERED 12 -> 10 on 2026-08-24 (batch 6): keepalive_outcomes (17 findings) and redirect_sweep (18) typed clean, 89 of 99 modules checked, 240 -> 205 findings. Both modules were drained by WRITE-MODE isolated offloads (dispatcher.offload --isolate) rather than in-seat, and every diff was reviewed here before it was applied -- the two changes that could have altered behaviour were checked first-person and both are equivalent: redirect_sweep's marker-rc read keeps the same except clause wrapping it, so a non-numeric rc still raises inside the try and still lands at None; keepalive_outcomes' added `oc is not None` guard short-circuits ahead of _should_record_outcome, whose own first line already returns False on oc is None. LOWERED 10 -> 7 on 2026-08-24 (batch 6b): dispatcher (65 findings), capability_propensity (47) and runtime_ac_gate (34) typed clean, 92 of 99 modules checked, 205 -> 91 findings. All three drained by write-mode isolated offloads and then INTEGRATED BY HAND rather than applied as returned -- three of the agents' choices were replaced with smaller ones, and the reasons are the durable part. (1) capability_propensity came back with two `type: ignore`s at the correlated-arm lookup, justified as an ambiguous key type. The ambiguity was real but the fix was in the wrong file: research_subjects.reciprocal_evidence_weights is GENERIC over its member type -- run-id strings at one caller, verdict indices at the other -- and its `dict[str, float]` return was too narrow. It is a TypeVar now and both ignores are gone. Suppressing there would have hidden a false positive AND blinded the one check that would catch a real key-type mismatch in the correlated-arm discount, which CLAUDE.md 2 makes load-bearing. (2) capability_propensity.detect() came back rewritten into four locals; behaviour-preserving (verified: same objects mutated, same key insertion order) but one `out: dict[str, Any]` annotation does the same job, so the annotation was taken instead. (3) runtime_ac_gate's spec_path came back with `spec_dir or env.get(K, DEFAULT)` split into branches, which silently falls back to DEFAULT where the original raised TypeError on Path(None). Unreachable via os.environ, but this is a GATE and 'silently use a default' is the permissive direction; with env typed as Mapping[str, Any] the original one-liner type-checks unchanged, so the restructure bought nothing. Asserts added in production paths were each checked against their invariant rather than trusted: dispatcher's `profile_attempt_id is not None` sits inside `elif selected_profile:` where line 815 makes it non-None by construction, and attach_profile_attempt_to_decision already declares `decision_id: str`. LOWERED 7 -> 1 on 2026-08-24 (batch 6c): capabilities (14 findings, ONE missing annotation on KNOWN_GATES caused all of them), router (16), and the last five findings across feedback/runtime_ac/capability_outcome_bridge/capability_compiler. 98 of 99 modules now checked; only capability_advisor (24) remains, held back because PR #113 edits the same file. The drain added ZERO net type: ignore -- router's one remaining suppression predates this work. THE DRAIN FOUND A REAL BUG and the value is in how it was handled: router's load_backlog promised a bare-list payload in its own docstring and could never read one, because `.get` on a list raises AttributeError into a broad `except` that returns [] -- so a populated backlog read as NO WORK. The offloaded agent was told to report suspected real bugs rather than fix them, did exactly that, and left a type: ignore documenting it; the fix landed here instead, with a selftest asserting a non-empty list survives (an empty one cannot distinguish the fix from the failure, since [] is also what a parse error returns) and a break->revert showing the old one-liner returning [] for a valid backlog. Latent, not live: the file on disk is the {\"items\": [...]} dict form, whose behaviour including its no-items fallback is unchanged. Worth closing anyway -- the day discovery writes the documented shape the symptom would be silence, which is this repo's signature failure. LOWERED 1 -> 0 on 2026-08-24 (batch 6d, THE LAST ONE): capability_advisor's 24 findings typed, the [[tool.mypy.overrides]] block removed entirely, and all 99 of 99 modules now checked. 240 findings drained to zero across the campaign with ZERO net `# type: ignore` added. THE LAST MODULE EXPOSED A LATCHED GATE IN THE RATCHET ITSELF, which is the durable finding here. verify.mypy_exempt_modules() returned None for BOTH 'pyproject unreadable' and 'readable, no override block' -- so the run that FINISHED the drain would have printed 'mypy ratchet: NOT COUNTED', the gate going silent at the exact moment it succeeded. Its own docstring already argued against this ('a ratchet that stops being counted is indistinguishable from one that emptied, and only one of those is good news'), and two more things prove it was unintended: _format_mypy_exempt_line carried a ' -- fully drained' branch no input could reach, and the selftest asserted the function was TRUTHY, so an empty list failed it. That is the CLAUDE.md pattern exactly -- a gate whose clear path is blocked by the thing it measures -- and it would have latched on the last module. Fixed: [] now means answered-and-empty, None stays unanswerable, and the selftest asserts all three renderings plus `is not None` rather than truthiness. `collected` is unchanged at 453: the new coverage lives in verify.py's own selftest, which pytest does not collect. FLOOR 500 -> 502 on 2026-08-25 by the capacity-arity branch: +2 tests, both pinning that `capacity.compute` returns THREE values on every branch. It used to return 2 or 3 depending on which of `_classify`'s 21 returns ran, and the first of those is the 429-shed check, whose `_shed` reads a file on the HOST outside `$ORCH_STATE_DIR` \u2014 so the ARITY was a property of the machine, and `test_capacity_gate_is_seat_level_not_gemini_special` died on the unpack locally while CI, where nothing is shed, stayed green. Measured on the merge result AFTER rebasing onto origin/main, which had taken #140 (500) mid-session: 502 on the rebase, not 460 on the pre-#140 branch, which is exactly the drift this equality exists to catch.",
- "mypy_exempt_max": 0
+ "mypy_exempt_max": 0,
+ "note": "Recorded by verify.py --update-floor, except the *_max ceilings, which are edited BY HAND and never re-measured. `collected` catches tests that stopped being collected; `passed` is compared against passed+skipped, so a check may move between passing and consciously-skipped but the two together may never shrink. The *_max ceilings bound the skipped side: 24/7/2 is exactly what a machine with none of this instance's local prerequisites skips (a GitHub runner: no agent CLIs, no ~/.codex/skills, no /Applications/ChatGPT.app, no populated capability ledger), measured 2026-08-21. On the owner's machine all prerequisites exist and nothing skips at all. Raising a ceiling is a deliberate act: it means agreeing that one more thing is allowed to go unchecked, so say which and why in the commit. LOWERED 26 -> 24 on 2026-08-22, reverting the raise made earlier the same day. The two kill-switch exemption tests no longer need to skip on a bare runner: their declarations moved out of the running instance's ledger and into capabilities.KNOWN_DECLARATIONS, so they assert code-derived truth and run everywhere. Moving a test back below the ceiling is the preferred way to lower it -- fix what made it machine-dependent, rather than agreeing to check less. FLOOR 345 -> 353 on 2026-08-22: 345 was measured on a branch cut before #13 (research panels/rounds/domain studies) merged, so the recorded floor sat 8 tests BELOW what main actually collects. A floor below reality is the permissive direction -- those 8 could have silently stopped being collected and still cleared the check, which is exactly the hole this file exists to close. Measure the floor on the merge result, not on the branch. Raised again on 2026-08-22 by the producer-identity-scope branch, which adds tests on top of the 353 recorded by #15; re-measured after rebasing rather than assumed. NOTE: `verify.py --update-floor` REPLACES this note with a generic one, so it must be restored by hand after every use \u2014 the ceiling rationale is the only record of which prerequisite justifies each skip. FLOOR 365 -> 366 on 2026-08-22 (heartbeat-ordering work, PR #18): exactly one new test, test_capabilities.test_no_tick_producer_runs_above_the_heartbeat_export. No ceiling moved and nothing new is skipped -- it reads source files rather than a populated ledger, so it runs on any machine. The branch recorded 354 because it was cut before #16 merged; re-measured on the MERGE RESULT per the rule above, which is exactly the mistake that put the floor 8 below reality last time. FLOOR 366 -> 368 on 2026-08-23: main collected 368 while this file recorded 366, drift left by #34 (evidence-acquisition landed, +1) and #37 (tick capability evidence, +1) whose authors each measured against a branch cut before the other merged. A floor BELOW reality is the permissive direction this file exists to close -- those two could have silently stopped being collected and still cleared the check. Measured on the merge result per the rule above: 368 passed, 0 failed, 0 skipped, 83/83 selftests, 43/43 can-fire, 5/5 gates. CEILING 24 -> 26 and FLOOR 368 -> 387 on 2026-08-23 (profiles/provenance branch, PR #42). This file CONFLICTED with #50, which raised the floor 366 -> 368 on main while this branch raised it to 387; resolved as the UNION rather than by taking a side -- #50's rationale is retained above and the count was RE-MEASURED on the new merge result instead of keeping either number. 368 (main) + 19 (this branch's net new tests) = 387; #50 corrected recorded drift rather than adding coverage, which is why 387 is unchanged from the pre-conflict measurement. Measured in a runner sandbox reproducing CI exactly (361 passed, 26 skipped, 387 collected) AND on the owner's machine (387 passed, 0 skipped, 5/5 gates). The two new skips are drift detectors against a REAL installed agent runtime, so neither can be moved below the ceiling -- the preferred way to lower one: (1) agy advertised-models cache absent, since comparing declared model ids against the catalogue agy actually advertises needs that catalogue, and a fixture would exercise the comparison while detecting no real drift; (2) vibe config absent (~/.vibe/config.toml), since active_model cannot be read to check for drift when there is no config to read. Both name their missing prerequisite, so a green run still states what it did not check. A third candidate skip was REFUSED: dispatcher's per-run agy-log assertion failed on a bare runner because adapters.advertised_models shells out to `agy models` when its disk cache is cold, and that probe landed inside a monkeypatched subprocess.run and overwrote the captured command. That is a stub leak, so it was fixed by ISOLATING the double rather than by skipping -- which makes CI run MORE. FLOOR 387 -> 391 on 2026-08-23 (improvement-log accessor, PR #59): exactly four new tests, all in test_improvement_log.py -- three read tracked files in the tree (the pointer's size and content, and that CLAUDE.md 0 step 3 and 5 name the accessor rather than a bare path) and one runs the accessor as a subprocess against a path that cannot exist. None reads a populated ledger, an agent CLI or ~/.codex, so all four RUN on a bare runner and NO ceiling moved: nothing new is skipped. Measured on the MERGE RESULT after rebasing onto origin/main af6654d, which collected 387 -- not on the branch base, per the rule above. FLOOR 391 -> 402 on 2026-08-23 (Gate python-ci configuration, the PR that adds the missing .github/workflows/autofix-versions.env): exactly 11 new tests, all in test_ci_gate_config.py, which read committed files only -- the pin file, ruff.toml, mypy.ini, pr-00-gate.yml's toggle annotations and docs/CI_LINT_BASELINE.md. NO ceiling moved. On any CHECKOUT -- CI, the owner's tree, a second instance -- all 11 run: they need no installed linter and no populated ledger. In the EXEC-MIRROR layout all 11 skip with one named reason, because orch-sync-mirror.sh copies root-level *.py only, so .github/workflows, docs/ and scripts/ are genuinely absent there (env_prereq.repo_files_absent). That lands at 11/26 on a machine that otherwise skips nothing, and CI stays at 26/26, so no ceiling needed raising. The skip gate is the presence of those DIRECTORIES, never of the pin file itself -- gating on the file would have made the test that checks for it unable to fail. Measured on the merge result, twice: the branch was rebuilt on origin/main after #42 and #59 merged, and re-measured after #61 merged and was merged in -- 393 passed + 9 skipped = 402 collected both times, so #61 added no collected tests and this floor is not sitting below reality. #61 itself left main's floor at 391, which is exactly main-without-these-11, so there is no inherited drift to correct. RULE CHANGE 2026-08-23: `collected` is now an EQUALITY, not a minimum. Every floor entry above this one records the number being found BELOW reality and hand-raised after the fact -- 21 low at the worst, then 8, then 1, then 2 -- because nothing ever required a test-adding PR to touch this file, so the permissive direction was silent by construction and the rule 'measure on the merge result' had to be restated three times with nothing enforcing it. verify.py now FAILS when collected exceeds the floor, printing the two integers to write. That also makes the concurrency case self-enforcing: once every test-adding branch must edit these same two lines, two concurrent branches CONFLICT IN GIT, so the second cannot merge without rebasing onto the first and re-measuring on the actual merge result. Demonstrated repeatedly on the change itself: six merges landed on main in the two hours it took to write, moving this file 368 -> 387 -> 391 -> 402, and every one would have left the floor below reality under the old one-directional rule. `passed` deliberately stays a MINIMUM on passed+skipped: only collection is machine-invariant (a skipped test is still collected), measured across machines at 391 collected on both, with pass/skip splits of 365/26 on CI against 391/0 locally. The *_max ceilings are untouched by this change and nothing new is skipped. `--update-floor` also stops REPLACING this note -- it appends -- so the warning above about restoring it by hand no longer applies; and drift does NOT block --update-floor, since a gate that forbade its own only remedy would be a deadlock (the first draft was exactly that). FLOOR 402 -> 407 on 2026-08-23 (findability admission requirement). (findability admission requirement). (findability admission requirement). (findability admission requirement). Exactly five new pytest tests, all in test_capability_admission.py: test_findability_distinguishes_its_three_sub_causes, test_findability_blocks_new_capabilities_and_reports_older_ones_as_debt, test_unreadable_reach_is_not_evaluated_and_never_a_failure, test_findability_exemption_is_declared_in_code_not_in_a_live_ledger, test_consult_sites_are_falsifiable_claims_about_real_callers. NO CEILING MOVED and nothing new skips: all five build synthetic ledgers in a tempdir or read committed tables, so none needs a populated capability ledger, an agent CLI or ~/.claude/skills. The one machine-dependent thing they touch -- an external consult site declared in capability_advisor.CONSULT_SITES whose skill prompt is not on this machine -- is reported as UNVERIFIED rather than skipped, because absence of the caller is not refutation of the claim; the in-tree site (tick) is asserted verified on every machine so the check can never degrade into 'everything unverified, nothing tested'. Measured on the merge result per the rule above: this file CONFLICTED three times while the branch was open, as main went 387 -> 391 -> 402 (#61, #64, #65, #60). Each time it was resolved as the UNION rather than by taking a side, and the count was RE-MEASURED on the new merge result rather than either number being carried forward: 402 (main at bd6da2e) + 5 (this branch's new tests) = 407. That is the rule this file already states -- measure the floor on the merge result, not on the branch -- and it mattered here, because #60 both deleted test_ci_gate_config.py and added more than it removed, so guessing in either direction would have been wrong. -> re-measured on 2026-08-23 (PR #62, the four deferred #42 review findings): three new tests, all machine-independent (each builds its own tmp_path Brain and manifests), so NO CEILING MOVED and nothing new is skipped. Fourth conflict for this branch, and the first one under the EQUALITY -- which is the point: the equality's own rationale says git conflict detection is what enforces 'measure on the merge result', and that is exactly what happened here. Under the old minimum the three earlier conflicts could each have been resolved by keeping the larger number; under the equality the count MUST be measured, and it was. RESOLVED AGAINST #68 (findability admission requirement) on 2026-08-23: taken as the UNION per the rule this file states -- #68's five-test entry is retained above and this branch's three-test entry beside it -- and the count RE-MEASURED on the merge result rather than keeping either side's number. main fc1fd42 collects 407; this branch adds 3; 410 measured with `pytest --collect-only -q` on the merge result, not assumed. Ceilings untouched at 26/7/2 and nothing new is skipped. Also resolved in the same merge: langsmith-fleet-worker-attempt.json, a CI-emitted `langsmith-fleet/v1` worker-attempt record whose two sides differed only in `emitted_at` and `pr_number` (62 here, 68 on main). Main's NEWER record was kept rather than this branch's older one -- discarding a newer provenance observation to win a merge would corrupt exactly the causal-provenance evidence CLAUDE.md 2 protects, and this branch's own run re-emits its record anyway. FLOOR 410 -> 411 on 2026-08-23 (CodeRabbit follow-up on PR #42, thread 3837879039; re-measured again after #56 made `collected` an EQUALITY, which makes an assumed number a hard RED rather than a quiet pass -- main stayed at 402 across #56, and the merge result measures 403, so #56 added no collected tests and this is main's 402 plus this branch's one): exactly one new test, test_feedback_model_provenance.test_late_sweep_completes_terminal_attempts_never_one_in_flight, which pins that ledger_reconcile.resolve_unresolved_worker_attempts completes only TERMINAL unresolved worker attempts and never one still in flight. No ceiling moved and nothing new is skipped -- the test builds its own tmp ledger and codex rollout fixture and monkeypatches adapters.CODEX_SESSIONS, so it needs no agent CLI and no populated capability ledger and runs on a bare runner. RESOLVED AGAINST #59 (improvement-log accessor), which raised the floor 387 -> 391 on main while this branch raised it to 388: taken as the UNION -- #59's rationale is retained above and the count was RE-MEASURED on the new merge result rather than keeping either number, which is the rule this file states and the mistake that once put the floor 8 below reality. 391 (main, incl. #59's four tests) + 1 (this branch's one new test) = 392 measured, not assumed: 392 passed, 0 failed, 0 skipped, 83/83 selftests, 43/43 can-fire, 5/5 gates. Three sibling follow-up branches are in flight against this same main (CI/ruff config, arm-attribution + durability, adapters label->ID); if this file conflicts with one of them, resolve as the UNION and RE-MEASURE on the new merge result rather than taking either number -- that is what #42 and #50 did, and taking a side is what put the floor 8 below reality earlier. RESOLVED AGAINST #68 (findability admission requirement) on 2026-08-23: taken as the UNION per the rule this file states -- #68's five-test entry is retained above and this branch's one-test entry beside it -- and the count RE-MEASURED on the merge result. main fc1fd42 collects 407; this branch adds 1; 408 measured with `pytest --collect-only -q` on the merge result, not assumed. Ceilings untouched at 26/7/2 and nothing new is skipped -- the one new test builds its own tmp ledger and codex rollout fixture, so it runs on a bare runner. Also resolved in the same merge: langsmith-fleet-worker-attempt.json, a CI-emitted `langsmith-fleet/v1` worker-attempt record differing only in `emitted_at` and `pr_number`; main's NEWER record was kept, since discarding a newer provenance observation to win a merge would corrupt the causal-provenance evidence CLAUDE.md 2 protects. FLOOR 411 -> 415 on 2026-08-23 (PR #70 diagnostics salvage): four new collected tests in test_capability_set_coverage.py from the PR #43 salvage plus CodeRabbit follow-ups on PR #51/#70 \u2014 union/missing-candidate fetch command, truncation after six modules, AST-scoped gate-call audit, and entrypoint-diagnosis coverage. NO CEILING MOVED and nothing new is skipped; all inject synthetic ledgers or read committed source. Measured on the merge result at 91d37fa: 389 passed + 26 skipped = 415 collected on CI, not assumed. FLOOR 415 -> 416 on 2026-08-23 (the dangling-citation follow-up, PR #74): exactly ONE new test, test_ci_gate_config.test_every_cited_repo_path_resolves, which reads the two committed config files this repo OWNS (the pin file and ruff.toml) and asserts every repo-relative path they cite exists. It exists because the pin file shipped citing docs/ci/LINT_BASELINE.md when the real path was docs/CI_LINT_BASELINE.md: the sibling checks read that file's CONTENTS thoroughly and its PROSE not at all, and the prose is the only pointer telling a reader where to re-measure before bumping a pin. Scoped to the two owned files deliberately -- scanning pr-00-gate.yml yields six findings that are all correct as written (guarded by hashFiles or a .agents check, or upstream paths), and a test that cries wolf gets waived. NO ceiling moved. RE-MEASURED SIX TIMES as the base moved under this ONE-LINE change: bd6da2e 402 -> ddb0928 402 -> fc1fd42 407 -> 0d661e3 407 -> 0593eeb 411 -> 6fed4ad 415, each +1 with this test, and the branch was rebuilt on each rather than the number carried forward. THIS BRANCH IS THE WORKED EXAMPLE of the equality's concurrency cost, so record it rather than rediscover it: main moved EIGHT times in the ~2.5 hours a one-line comment fix was open (#56, #68, #73, #69, #62, #70 and two direct commits), the floor line conflicted THREE separate times, and two merges overlapped the change directly -- #73 landed a byte-identical copy of the backplane-conformance.yml guard this branch also carried (dropped as redundant), and #69 edited this very test file in a neighbouring region. The equality is still the right call and should stay: every entry above this one records the floor being found BELOW reality, which is the permissive direction. But no amount of author care wins this race, because the correct value is only knowable on the merge result. The durable fix is CI running `verify.py --update-floor` on the merge commit, which keeps the equality and removes the race; until then a test-adding PR must be merged promptly after going green, because it re-conflicts on roughly every subsequent merge. FLOOR 416 -> 427 on 2026-08-23 (PR #72 hygiene untrack, rebased after #71 merged): exactly 11 new tests from test_repo_artifact_hygiene.py with root-anchored gitignore patterns. NO ceiling moved. Measured on merge result after #71 landed on main: 416 (main) + 11 = 427 collected via pytest --collect-only -q, not assumed. #71's simpler untrack landed first; this branch carries the full hygiene test suite and corrected root-anchored patterns. FLOOR 427 -> 428 on 2026-08-23 (PR salvaging #34/#42 remnants): exactly one new test, test_feedback_model_provenance.test_gemini_provenance_reads_the_per_run_log_before_the_conversation_store, recovered from #42's post-merge commit 4e0d6ae along with the adapters catalog work it exercises. No ceiling moved and nothing new is skipped -- it seeds adapters._ADVERTISED_MEMO instead of letting the catalog probe shell out, so it runs on any machine and adds no prerequisite. `passed` is 428 rather than the 426 verify.py suggested on this machine: two test_capabilities liveness tests (test_gate_blocks_execution_is_opt_in_and_narrow, test_evidence_gate_kind_is_not_blanket_observer) currently fail HERE on pristine main as well, because the hourly fleet tick mutated the machine-local ledger and range-lane-rollout now classifies matched_not_invoked rather than deliberately_gated. That is ledger STATE, not this branch and not the code -- CI bootstraps an empty ledger and counts 428/428. Recording 426 would have baked a local environment failure into the floor as though it were the expected result. FLOOR 428 -> 441 on 2026-08-23 (coverage measures what actually runs): exactly 12 new tests, all in test_verify_coverage_mode.py. They read committed files and verify.py's own source, and monkeypatch verify.COVERAGE in-process -- no populated ledger, no agent CLI, no ~/.codex, and no coverage RUN -- so all 12 execute on any machine and NO ceiling moved. The change itself is a measurement fix, not a gate: `verify.py --coverage` wraps each child in `coverage run --parallel-mode` and combines, because the per-module --selftest is a SUBPROCESS and a pytest-only coverage run cannot see it. That blind spot was most of the codebase -- 78 modules have no test_*.py at all, ~85,500 lines, 79.6% of non-test root Python -- so the reported 48.45% was measuring the gap in the instrument, not a gap in the tests. Combined: 76.1% (45,049 statements, 10,774 missed). Twelve of the twelve modules the old report named as worst were selftest-only; outcomes.py reported 9.0% and measures 61.9%, watch.py 9.6% -> 86.9%. Coverage is OFF by default and deliberately never touches the exit code -- one of the 12 tests pins that, because enforcing a threshold here would reward pytest wrappers around already-tested modules: metric up, assurance flat. Measured on the MERGE RESULT: branched from origin/main bcc68cd (floor 427), then REBASED onto f5f1c39 when it landed underneath and re-measured on the new merge result rather than carrying the old number: 428 + 13 = 441. +1 on 2026-08-23 (440 -> 441 after the rebase), same branch: test_the_cli_help_actually_renders. It exists because this branch BROKE `verify.py --help` and its own twelve tests did not notice. argparse interpolates help strings with `% params`, so the literal `~80%` in the --coverage help was read as an `%o` octal conversion and --help died with 'badly formed help string'. All twelve original tests passed: every one inspected source text or monkeypatched a flag, and not one RENDERED the help -- a construction-time test suite that never exercised the constructed thing, which is this repo's founding defect one layer up. CI's verify.py gate caught it, which is the check of last resort working as intended. The new test runs `verify.py --help` as a subprocess and asserts it exits 0, so it runs anywhere and NO ceiling moved. Rendering rather than grepping for `%` is deliberate: a grep would flag the legitimate `%(default)s`. FLOOR 441 -> 442 on 2026-08-23 (matched_not_invoked yields to observers and declared gates): exactly one new test, test_capabilities.test_matched_not_invoked_yields_to_observers_and_declared_gates. It is SYNTHETIC on purpose and that is the point of it: the two tests that caught this bug in the wild read the LIVE ledger, so they skipped with a named reason on the empty ledger ci.yml bootstraps -- the defect was red on every populated machine and green on CI for as long as it existed. A synthetic row asks the same question everywhere, so this one RUNS on a bare runner and NO ceiling moved. The fix itself moves matched_not_invoked below `observing` and below the DECLARED deliberately_gated check in classify_liveness: it was the first check, which made it the fourth instance of the unescapable label the comments in that function exist to fix. Audited before committing -- 12 of 43 live rows reclassify (ten observers to observing, two declared gates to deliberately_gated) and ZERO move for any other reason, so nothing is reclassified by inference and the weaker gate_reason-only branch is untouched. Measured on origin/main 0d9c3a7, whose recorded floor is 441, so 441 + 1 = 442. FLOOR RECORDED by verify.py --update-floor on 2026-08-23: collected=442, passed=442. Ceilings preserved, never re-measured \u2014 they are edited by hand. FLOOR RECORDED by verify.py --update-floor on 2026-08-24: collected=442, passed=442. Ceilings preserved, never re-measured \u2014 they are edited by hand. FLOOR RECORDED by verify.py --update-floor on 2026-08-24: collected=442, passed=442. Ceilings preserved, never re-measured \u2014 they are edited by hand. MYPY_EXEMPT_MAX introduced at 64 on 2026-08-23, the same change that flipped `typecheck` ON in pr-00-gate.yml. It was OFF because 608 whole-tree errors were drainable 0 per PR -- a gate whose clear path is blocked by the thing it measures. Two real changes opened it: the src/ move scoped the Gate's `target=src` to the 99 modules (608 -> 467), and pyproject.toml's [[tool.mypy.overrides]] exempts the modules that still have findings BY NAME so the 35 already-clean ones are checked today. 12 var-annotated findings were then drained (467 -> 455, 66 -> 64 modules) to prove the drain works rather than promise it. This ceiling may only ever be LOWERED, by typing a module and deleting its line -- raising it means agreeing one more module goes unchecked, so say which and why. NOT a suppression: no error code is disabled anywhere and the 455 stay visible via `python3 scripts/ci_lint_baseline.py`. Edited BY HAND like the other *_max values; --update-floor never re-measures it. MYPY_EXEMPT_MAX LOWERED 64 -> 43 on 2026-08-24, the ratchet's first real drain: 21 modules typed clean and removed from the exempt list, 455 -> 430 findings. Targeted the modules with 1-3 errors on purpose -- fixing 50 errors spread across the big modules would move this number by ZERO, and this number is what the gate reads. Lowering it is the drain; it may never be raised without naming the module and why. MYPY_EXEMPT_MAX LOWERED 43 -> 33 on 2026-08-24 (batch 2 of the drain): the whole <=5-finding tail typed clean. 66 of 99 modules now checked. What remains is five per-module campaigns (capability_advisor, dispatcher, capability_propensity, runtime_ac_gate, capabilities hold most of it), so future batches take ONE big module at a time rather than skimming. (batch 2 detail: 43 -> 26, 430 -> 374 findings, 73 of 99 modules checked. `mypy_path` gained `tests` so mypy can RESOLVE the recurrence-fixture roster capability_admission legitimately imports; the target stays `src`, but mypy then follows into that one test file, whose two findings were fixed rather than configured around.) MYPY_EXEMPT_MAX LOWERED 26 -> 20 on 2026-08-24 (batch 3): 79 of 99 modules checked. FLOOR 442 -> 448 on 2026-08-24 (absent-check detector + its ratchet): SIX new tests in tests/test_checks_reported.py holding the frequency rule and the expected-check ratchet. NO ceiling moved -- they call pure functions and read one committed JSON file, so they run on any machine. The ratchet exists because dogfooding the detector caught it DISARMING ITSELF: while pr-00-gate.yml sat held, every merged PR merged without the Gate, so after twelve such merges the Gate's checks fell below the 75% frequency threshold, the expected set eroded 23 -> 14, and PR #91 was pronounced clean by the tool written to catch exactly that. config/expected-checks.json is the high-water mark, seeded from PRs #87/#89 whose Gate demonstrably ran, and it only comes down when somebody deletes a line. LOWERED to 16 on 2026-08-24 (batch 4, re-applied after merging main, which carried #94/#100's own floor work \u2014 main's note kept, only the bound re-set). LOWERED 16 -> 16 on 2026-08-24 (batch 5): 83 of 99 checked. LOWERED 16 -> 12 on 2026-08-24 (batch 5): 87 of 99 checked. Lesson recorded three times now: measure with the PROJECT run, never per-file \u2014 `mypy src/X.py` reports clean for modules the project run still flags. FLOOR 448 -> 450 on 2026-08-24 (giving the per-node deliberate-break finding its consumers): exactly two new pytest tests, both in test_synthesis_promotion.py -- test_a_passing_break_names_the_tautologies_it_carried and test_the_break_caveat_is_silent_on_clean_and_on_pre_per_node_evidence. NO CEILING MOVED and nothing new is skipped: the first builds its own git repo in tmp_path and runs the real verification path (git + pytest only, no populated ledger, no agent CLI, no ~/.codex), and the second is a pure unit on _break_caveat. The third piece of this change is a runtime_ac selftest case, which adds NO collected test -- runtime_ac has a --selftest and synthesis_promotion does not, which is the whole reason only one of the two needed pytest tests. Measured on the merge result: branched from origin/main 8549f84, whose recorded floor is 448, and re-checked that main had not moved before writing this. 448 + 2 = 450, from verify.py's own count (450 passed, 0 failed, 0 skipped, 85/85 selftests, 5/5 gates), not assumed. FLOOR 448 -> 451 on 2026-08-24 (offload guidance correction, PR #113): THREE new tests in test_repo_artifact_hygiene.py -- two parametrized entries adding src/UNKNOWN.egg-info/PKG-INFO and UNKNOWN.egg-info/PKG-INFO to EMITTED_ARTIFACTS, and test_no_build_metadata_is_tracked. NO ceiling moved and nothing new skips: all three shell out to git against this checkout, so they run anywhere git does. They exist because THIS PR carried four committed build artifacts and no check objected. A CI step runs `pip install -e .`, setuptools writes src/UNKNOWN.egg-info/ (named UNKNOWN because pyproject.toml declares no [project] on purpose), and the repo's own autofix bot committed all four -- .gitignore had no *.egg-info/ entry and this suite only knew the langsmith-fleet family. CodeRabbit caught it; nothing local did. Both locations are listed because a checkout builds into src/ while the EXEC MIRROR IS FLAT. The tracked test is separate from the ignored test on purpose: the four files were committed BEFORE the pattern existed, and adding a pattern does nothing to a path git already tracks -- so an ignore-only check passes on a repo still carrying the debris. RESOLVED AGAINST #110 and #114 on 2026-08-24, which moved main 448 -> 450 while this branch raised it to 451: taken as the UNION per the rule this file states -- main's entries are retained above and this branch's three-test entry beside them -- and the count RE-MEASURED on the merge result rather than either number being carried forward. 450 (main) + 3 (this branch) = 453, measured with `pytest --collect-only -q`, not assumed. Ceilings untouched and nothing new is skipped. FLOOR 453 -> 458 on 2026-08-24 (untracking .coverage, the coverage database this repo's own verify.py writes): exactly FIVE new tests, all in test_repo_artifact_hygiene.py -- two parametrized ignore cases over COVERAGE_DATA_FILES (.coverage and a parallel-mode .coverage.<host>.<pid>.<random>), one untracked case, and two parametrized must-stay-committable cases (.coveragerc, tools/coverage_guard.py). NO CEILING MOVED and nothing new is skipped: all five ask GIT about a path in this checkout (check-ignore / ls-files) and need no populated ledger, no agent CLI and no ~/.codex, so all five RUN on a bare runner. The change is `git rm --cached .coverage` plus root-anchored `/.coverage` and `/.coverage.*`; the 90 KB SQLite database arrived on main in #109, a typing PR whose every other file is about mypy, and verify.py's coverage_reset() unlinks it and coverage_combine_and_report() rewrites it on every --coverage run -- so while tracked it was an opaque binary rewritten by the command that produces this repo's verdict. Both patterns because two different steps write them, and `/.coverage.*` rather than `/.coverage*` because the second swallows .coveragerc. Break -> revert performed in all three directions and recorded in the test file. Measured on the MERGE RESULT: fast-forwarded onto origin/main 5c769e0, whose recorded floor is 453, and re-fetched to confirm main had not moved again before measuring. 453 + 5 = 458 from verify.py's own count (458 passed, 0 failed, 0 skipped, 85/85 selftests, 5/5 gates), not assumed. LOWERED 12 -> 10 on 2026-08-24 (batch 6): keepalive_outcomes (17 findings) and redirect_sweep (18) typed clean, 89 of 99 modules checked, 240 -> 205 findings. Both modules were drained by WRITE-MODE isolated offloads (dispatcher.offload --isolate) rather than in-seat, and every diff was reviewed here before it was applied -- the two changes that could have altered behaviour were checked first-person and both are equivalent: redirect_sweep's marker-rc read keeps the same except clause wrapping it, so a non-numeric rc still raises inside the try and still lands at None; keepalive_outcomes' added `oc is not None` guard short-circuits ahead of _should_record_outcome, whose own first line already returns False on oc is None. LOWERED 10 -> 7 on 2026-08-24 (batch 6b): dispatcher (65 findings), capability_propensity (47) and runtime_ac_gate (34) typed clean, 92 of 99 modules checked, 205 -> 91 findings. All three drained by write-mode isolated offloads and then INTEGRATED BY HAND rather than applied as returned -- three of the agents' choices were replaced with smaller ones, and the reasons are the durable part. (1) capability_propensity came back with two `type: ignore`s at the correlated-arm lookup, justified as an ambiguous key type. The ambiguity was real but the fix was in the wrong file: research_subjects.reciprocal_evidence_weights is GENERIC over its member type -- run-id strings at one caller, verdict indices at the other -- and its `dict[str, float]` return was too narrow. It is a TypeVar now and both ignores are gone. Suppressing there would have hidden a false positive AND blinded the one check that would catch a real key-type mismatch in the correlated-arm discount, which CLAUDE.md 2 makes load-bearing. (2) capability_propensity.detect() came back rewritten into four locals; behaviour-preserving (verified: same objects mutated, same key insertion order) but one `out: dict[str, Any]` annotation does the same job, so the annotation was taken instead. (3) runtime_ac_gate's spec_path came back with `spec_dir or env.get(K, DEFAULT)` split into branches, which silently falls back to DEFAULT where the original raised TypeError on Path(None). Unreachable via os.environ, but this is a GATE and 'silently use a default' is the permissive direction; with env typed as Mapping[str, Any] the original one-liner type-checks unchanged, so the restructure bought nothing. Asserts added in production paths were each checked against their invariant rather than trusted: dispatcher's `profile_attempt_id is not None` sits inside `elif selected_profile:` where line 815 makes it non-None by construction, and attach_profile_attempt_to_decision already declares `decision_id: str`. LOWERED 7 -> 1 on 2026-08-24 (batch 6c): capabilities (14 findings, ONE missing annotation on KNOWN_GATES caused all of them), router (16), and the last five findings across feedback/runtime_ac/capability_outcome_bridge/capability_compiler. 98 of 99 modules now checked; only capability_advisor (24) remains, held back because PR #113 edits the same file. The drain added ZERO net type: ignore -- router's one remaining suppression predates this work. THE DRAIN FOUND A REAL BUG and the value is in how it was handled: router's load_backlog promised a bare-list payload in its own docstring and could never read one, because `.get` on a list raises AttributeError into a broad `except` that returns [] -- so a populated backlog read as NO WORK. The offloaded agent was told to report suspected real bugs rather than fix them, did exactly that, and left a type: ignore documenting it; the fix landed here instead, with a selftest asserting a non-empty list survives (an empty one cannot distinguish the fix from the failure, since [] is also what a parse error returns) and a break->revert showing the old one-liner returning [] for a valid backlog. Latent, not live: the file on disk is the {\"items\": [...]} dict form, whose behaviour including its no-items fallback is unchanged. Worth closing anyway -- the day discovery writes the documented shape the symptom would be silence, which is this repo's signature failure. LOWERED 1 -> 0 on 2026-08-24 (batch 6d, THE LAST ONE): capability_advisor's 24 findings typed, the [[tool.mypy.overrides]] block removed entirely, and all 99 of 99 modules now checked. 240 findings drained to zero across the campaign with ZERO net `# type: ignore` added. THE LAST MODULE EXPOSED A LATCHED GATE IN THE RATCHET ITSELF, which is the durable finding here. verify.mypy_exempt_modules() returned None for BOTH 'pyproject unreadable' and 'readable, no override block' -- so the run that FINISHED the drain would have printed 'mypy ratchet: NOT COUNTED', the gate going silent at the exact moment it succeeded. Its own docstring already argued against this ('a ratchet that stops being counted is indistinguishable from one that emptied, and only one of those is good news'), and two more things prove it was unintended: _format_mypy_exempt_line carried a ' -- fully drained' branch no input could reach, and the selftest asserted the function was TRUTHY, so an empty list failed it. That is the CLAUDE.md pattern exactly -- a gate whose clear path is blocked by the thing it measures -- and it would have latched on the last module. Fixed: [] now means answered-and-empty, None stays unanswerable, and the selftest asserts all three renderings plus `is not None` rather than truthiness. `collected` is unchanged at 453: the new coverage lives in verify.py's own selftest, which pytest does not collect. FLOOR 500 -> 502 on 2026-08-25 by the capacity-arity branch: +2 tests, both pinning that `capacity.compute` returns THREE values on every branch. It used to return 2 or 3 depending on which of `_classify`'s 21 returns ran, and the first of those is the 429-shed check, whose `_shed` reads a file on the HOST outside `$ORCH_STATE_DIR` \u2014 so the ARITY was a property of the machine, and `test_capacity_gate_is_seat_level_not_gemini_special` died on the unpack locally while CI, where nothing is shed, stayed green. Measured on the merge result AFTER rebasing onto origin/main, which had taken #140 (500) mid-session: 502 on the rebase, not 460 on the pre-#140 branch, which is exactly the drift this equality exists to catch. MIRROR CEILING ADDED 2026-08-29: `mirror_skipped_max` = 31. `skipped_max` was ONE number bounding TWO different deprived environments, and this note above says which one it was measured in -- a GitHub runner with none of this instance's local prerequisites. The EXEC MIRROR is a different shape: every local prerequisite is present there (it exists only on the machine the system runs on, so nothing skips for a missing CLI, skill, app bundle or ledger row) but it is a FLAT FILE COPY, so it is not a git repository and has no .github/. Measured on unmodified main: 471 passed, 31 skipped -- 12 from `repo_files_absent('.github/workflows', ...)` in test_ci_gate_config.py and 19 from `git_repo_absent()` in test_repo_artifact_hygiene.py, and ZERO runner-shape skips. Against the 26 measured on a runner that skips neither family, `python3 verify.py` from ~/.codex/orchestrator-mirror was RED ON EVERY INPUT, including an entirely correct tree -- confirmed structural by syncing the PREVIOUS main into a scratch mirror and getting the same 31. CLAUDE.md 1 makes the mirror run the verdict ('cmp-clean is not agreement'), so the one instrument that catches cross-tree divergence had stopped protecting anything, and a gate red whatever you do is a gate that gets switched off. NOT FIXED BY RAISING `skipped_max` TO 31: CI runs on a runner where 26 is the correct bound, and the extra five would have let a runner skip five more checks in silence -- one gate going quiet to un-stick another. Instead each shape carries its own agreed number, selected by `env_prereq.exec_mirror_shape()`, which detects the PREREQUISITE (no .github/ AND git reports no repository) and never `$CI`, exactly as every other detector in that module does. Both marks are required because the mirror number is the LOOSER one, so an ambiguous tree falls back to the base. The mirror value REPLACES the base rather than adding to it: the runner-shape absences the base pays for cannot occur on the mirror, and 26+31 would have bought 26 units of headroom nothing there can legitimately spend. WHAT DRAINS IT: making a mirror-skipped test runnable there. Teaching orch-sync-mirror.sh to copy .github/ would drop 12 and this number must come down to 19 IN THE SAME CHANGE -- but note that script lives outside the repo, so a mirror synced by an unpatched copy would then be red at 31 > 19, which is the disease this entry cures; do it only if the script's state can be relied on. The 19 git-shape skips are irreducible while the mirror is a file copy. Any ceiling key may carry a `mirror_` variant by the same derived rule; only this one is set, and an unset variant falls back to the base agreement, which is the strict direction."
 }
diff --git a/CLAUDE.md b/CLAUDE.md
index fea495d..73546a5 100644
--- a/CLAUDE.md
+++ b/CLAUDE.md
@@ -275,6 +275,20 @@ Do not create a second event log, model registry, or capability inventory.
   that makes CI run MORE. **Never turn a real failure into a skip**, and never add a skip without
   a reason string — a reason-less skip is indistinguishable from a pass, which is this repo's
   founding defect wearing a different hat.
+- **A CEILING BOUNDS ONE DEPRIVED SHAPE, AND THERE ARE TWO.** `skipped_max` was measured on a bare
+  GitHub runner — no agent CLIs, no `~/.codex/skills`, no app bundle, no populated ledger — and was
+  then also applied to the EXEC MIRROR, which is the opposite deprivation: every local prerequisite
+  present, but a flat file copy with no `.github/` and no `.git`, so it skips 31 tests a runner
+  skips none of. One number over two populations is the latched-gate defect in its purest form, and
+  the symptom was total: `python3 verify.py` from `~/.codex/orchestrator-mirror` was RED on every
+  input including a correct tree, for as long as it existed — while §1 above makes that run the
+  verdict. Raising the number would have been the wrong repair; it would have handed the RUNNER
+  five units of slack, where 26 is the measured bound. **Each shape carries its own agreed number,
+  measured where it is enforced**: `env_prereq.exec_mirror_shape()` detects the tree (both marks
+  required, `$CI` never consulted) and `verify.mirror_key` derives the floor key, so any ceiling may
+  carry a `mirror_` variant and an unset one falls back to the base — the strict direction. The
+  summary always prints which tree it decided it was in. When you add a ceiling, ask which shape you
+  measured it in before writing the number down.
 - **State lives behind TWO variables and they are not the same.** `ORCH_STATE_DIR` holds the audit
   cache, firing-monitor and redirect-sweep state; `ORCH_LOCAL_RUNTIME` holds the capability LEDGER
   and the Brain. Pointing only the first at an empty directory and concluding "the suite is
diff --git a/docs/MIRROR_SYNC_PATCH.md b/docs/MIRROR_SYNC_PATCH.md
index a8eecae..f7d9a2c 100644
--- a/docs/MIRROR_SYNC_PATCH.md
+++ b/docs/MIRROR_SYNC_PATCH.md
@@ -77,3 +77,15 @@ bash ~/.codex/bin/orch-sync-mirror.sh && ls ~/.codex/orchestrator-mirror/*.py |
 
 Expect ~99, and then `cd ~/.codex/orchestrator-mirror && python3 verify.py` should be green — note
 that in the FLAT mirror the command keeps its old form, with no `src/` prefix.
+
+**What "green in the mirror" means, and why it was impossible until 2026-08-29.** The mirror is a
+flat copy, so it is not a git repository and has no `.github/`; 31 tests skip there for exactly
+those two named reasons, and the run reports `31/31 max [mirror_skipped_max]`. Those skips are
+agreed, not tolerated — `.verify-floor.json` now carries a `mirror_skipped_max` alongside
+`skipped_max`, because the latter was measured on a bare GitHub runner (a different deprivation:
+missing CLIs and ledger rows, but a real checkout) and one number cannot bound both populations.
+A mirror run was therefore RED on every input, correct trees included, which made the instrument
+this doc points you at worthless. The shape is detected by `env_prereq.exec_mirror_shape()`, never
+from `$CI`, and the summary's first line says which tree it decided it was in. If you ever teach
+the sync to copy `.github/`, 12 of those skips become real checks and `mirror_skipped_max` must
+come down to 19 in the same change.
diff --git a/src/env_prereq.py b/src/env_prereq.py
index 7a70fd9..a3122dd 100644
--- a/src/env_prereq.py
+++ b/src/env_prereq.py
@@ -277,6 +277,59 @@ def git_repo_absent() -> str | None:
     return None
 
 
+def exec_mirror_shape() -> str | None:
+    """Name what makes THIS tree the exec mirror rather than a checkout, or None for a checkout.
+
+    NOT a prerequisite detector, and it must never be handed to `require()`. Every other function
+    in this module answers "can this CHECK run here", and its reason SKIPS something. This one
+    answers "which TREE is this", and its reason selects which agreed skip ceiling `verify.py`
+    enforces. Nothing skips because of it.
+
+    WHY IT EXISTS. `.verify-floor.json`'s `skipped_max` was ONE number bounding TWO different
+    deprived environments, and the floor file's own note says which one it was measured in:
+    "exactly what a machine with none of this instance's local prerequisites skips (a GitHub
+    runner: no agent CLIs, no ~/.codex/skills, no /Applications/ChatGPT.app, no populated
+    capability ledger)". The exec mirror is a DIFFERENT shape with DIFFERENT absences. Every local
+    prerequisite is present there -- it exists only on the machine the system runs on -- but it is
+    a flat file copy, so it is not a git repository and has no `.github/`. Measured 2026-08-29 on
+    unmodified main: the mirror skips the two families below, 31 tests, against a ceiling of 26
+    measured on a runner that skips neither, and both control runs (live mirror and a scratch
+    mirror synced from the previous main) gave the same 31. So `python3 verify.py` from the mirror
+    was RED on every input INCLUDING a correct tree -- and CLAUDE.md 1 makes that run the verdict,
+    so the one instrument that catches cross-tree divergence had stopped protecting anything. A
+    gate red on every input is a gate that gets ignored.
+
+    Raising `skipped_max` to 31 would have been the wrong repair: CI runs on a runner where 26 is
+    the correct bound, so five more checks could have gone unchecked THERE in silence. Two shapes
+    need two agreements, each measured in the shape it bounds -- the measuring window equal to the
+    draining window, which is the latched-gate rule this file already follows everywhere else.
+
+    ONE POPULATION, CAUSED AND LICENSED BY THE SAME TWO DETECTORS. The marks below are the very
+    calls that PRODUCE the mirror-only skips: `repo_files_absent` from `test_ci_gate_config.py`
+    and `git_repo_absent` from `test_repo_artifact_hygiene.py`. The ceiling that licenses those
+    skips is therefore unlocked by exactly the facts that cause them, with no second literal to
+    drift. It selects WHICH agreed number applies -- never what the number is, which stays
+    hand-edited -- so this cannot become a gate that clears itself.
+
+    BOTH marks are required, and that AND is the conservative direction: the mirror ceiling is the
+    LOOSER of the two, so an ambiguous tree must fall back to the base agreement rather than help
+    itself to headroom. A checkout with `git` uninstalled still has `.github/`; a checkout whose
+    `.github/` was deleted is still a git repository. Neither reads as the mirror.
+    """
+    no_github = repo_files_absent(".github")
+    not_a_repo = git_repo_absent()
+    if not (no_github and not_a_repo):
+        return None
+    import paths
+
+    here = paths.checkout_root(Path(__file__).resolve().parent)
+    return (
+        f"{here} is the exec mirror, not a checkout: no .github/ directory, and git reports no "
+        f"repository here. orch-sync-mirror.sh copies files, not a checkout, so the tests that "
+        f"read repository configuration or ask git a question cannot run in this tree"
+    )
+
+
 def codex_profile_binary_absent() -> str | None:
     """Reason string when the version-capable Codex binary exact profiles require is absent.
 
@@ -505,6 +558,52 @@ def _selftest() -> None:
         "pyproject.toml", "definitely-not-a-file-here"
     ), "one missing entry is enough to skip"
 
+    # ---- exec_mirror_shape: WHICH TREE, not "can this check run" ------------------------------
+    # The composition rule is the whole function, so it is asserted directly in all four
+    # combinations. AND, never OR: the mirror ceiling is the LOOSER of the two agreements, so only
+    # an unambiguous flat copy may unlock it. Flipping the `and` to `or` turns three of these red.
+    _real_github, _real_git = repo_files_absent, git_repo_absent
+    try:
+        for _no_github, _no_git, _want_mirror in (
+            (True, True, True),
+            (True, False, False),
+            (False, True, False),
+            (False, False, False),
+        ):
+            globals()["repo_files_absent"] = lambda *a, _r=_no_github: (
+                "no .github here" if _r else None
+            )
+            globals()["git_repo_absent"] = lambda _r=_no_git: (
+                "not a git repository" if _r else None
+            )
+            got_shape = exec_mirror_shape()
+            assert bool(got_shape) is _want_mirror, (_no_github, _no_git, got_shape)
+            if _want_mirror:
+                # It must NAME the tree and both marks, or a reader cannot tell why the looser
+                # ceiling is in force -- and an unexplained looser ceiling is the thing to fear.
+                assert got_shape is not None
+                assert ".github" in got_shape and "git" in got_shape, got_shape
+    finally:
+        globals()["repo_files_absent"] = _real_github
+        globals()["git_repo_absent"] = _real_git
+
+    # THE REAL TREE, asserted in the direction that is layout-independent AND permissive-side:
+    # a genuine checkout must never be mistaken for the mirror, because that is the only error
+    # that hands a real repository the bigger skip ceiling. The mirror direction is asserted too,
+    # from the same two facts -- and this pair is what running verify FROM THE MIRROR exercises.
+    import paths
+
+    _root = paths.checkout_root(Path(__file__).resolve().parent)
+    if (_root / ".github").is_dir():
+        assert exec_mirror_shape() is None, (
+            f"{_root} has .github/ and must read as a checkout, not the exec mirror: "
+            f"{exec_mirror_shape()}"
+        )
+    elif git_repo_absent():
+        assert (
+            exec_mirror_shape()
+        ), f"{_root} is a flat non-repository copy and must read as the mirror"
+
     # A skipped selftest must SPEAK, and its line must carry the shared mark verify.py greps
     # for. A skip that prints nothing is a silent zero-exit by another name.
     import contextlib
@@ -557,7 +656,7 @@ def _selftest() -> None:
 
     print(
         "env_prereq.py selftest: OK (skip-is-a-skip, every detector names the missing thing, "
-        "marked selftest skip speaks, vibe readers)"
+        "marked selftest skip speaks, vibe readers, exec-mirror shape needs BOTH marks)"
     )
 
 
@@ -566,6 +665,14 @@ def main(argv: list[str]) -> int:
         _selftest()
         return 0
     # Default: report what this machine can and cannot check. Useful on a new box.
+    # WHICH TREE first, because it changes which skip ceiling verify.py enforces below. Printed
+    # in both shapes: "checkout" is the answer as much as the mirror reason is, and a shape line
+    # that only appears in one of them tells a reader nothing on a machine where it is silent.
+    shape = exec_mirror_shape()
+    print(
+        f"tree shape: {'EXEC MIRROR' if shape else 'checkout'}"
+        + (f"\n    {shape}" if shape else "")
+    )
     checks = {
         "reference skill resource": skill_resource_absent(),
         "exact-profile Codex binary": codex_profile_binary_absent(),
diff --git a/src/verify.py b/src/verify.py
index 42c7582..9d6d6c6 100644
--- a/src/verify.py
+++ b/src/verify.py
@@ -40,9 +40,17 @@
     The floor is now `passed + skipped >= floor.passed`, not `passed >= floor.passed`: a check
     may move between passing and consciously-skipped, but the two together may never shrink.
     Turning a failure into a pass by skipping it is what the ceiling forbids; letting a machine
-    without the prerequisite report honestly is what the floor change allows. One constant per
-    ceiling, defined once in the floor file and read once here, so the measuring and draining
-    windows cannot drift apart.
+    without the prerequisite report honestly is what the floor change allows.
+
+    ONE CONSTANT PER CEILING **PER TREE SHAPE** (2026-08-29). "One constant, so the measuring and
+    draining windows cannot drift apart" was the rule and it was being broken by this file's own
+    ceiling: `skipped_max` was measured on a GitHub runner and then applied to the EXEC MIRROR
+    too, which is a different deprived shape — every local prerequisite present, but a flat file
+    copy with no `.github/` and no `.git`, skipping 31 tests a runner does not skip at all. So a
+    mirror run was RED on every input including a correct tree, while CLAUDE.md §1 makes that run
+    the verdict. Two shapes, two agreements, each measured where it is enforced: the shape is
+    detected by `env_prereq.exec_mirror_shape()` (the prerequisite, never `$CI`) and selects the
+    key via `ceiling_limit`. The summary always prints which tree it decided it was in.
 """
 
 from __future__ import annotations
@@ -140,10 +148,21 @@ def coverage_combine_and_report() -> str:
 # Imported from env_prereq rather than duplicated: a shared literal in two files is a pair that
 # drifts, and a mark that drifts turns a skip back into a silent pass.
 try:
-    from env_prereq import PREREQ_ABSENT_MARK
+    from env_prereq import PREREQ_ABSENT_MARK, exec_mirror_shape
 except Exception:  # noqa: BLE001
     PREREQ_ABSENT_MARK = "PREREQUISITE ABSENT:"
 
+    def exec_mirror_shape() -> str | None:
+        """Fallback when env_prereq will not import: read the tree as a CHECKOUT.
+
+        The strict direction on purpose. "checkout" selects the SMALLER agreed ceiling, so a tree
+        whose shape could not be determined is bounded by the tighter agreement and goes red if it
+        skips more. Defaulting to the mirror would hand an unknown tree the extra headroom, which
+        is the one error worth engineering against here.
+        """
+        return None
+
+
 # pytest's terse summary line, e.g. "182 passed, 3 skipped in 41.20s"
 COUNT_RE = re.compile(r"(\d+) (passed|failed|error|errors|skipped|xfailed|xpassed)")
 
@@ -542,6 +561,55 @@ def _appended_note(prior: str | None, collected: int, passed: int, today: str) -
 )
 
 
+# ONE NUMBER MAY NOT BOUND TWO DIFFERENT DEPRIVED ENVIRONMENTS, and until 2026-08-29 `skipped_max`
+# did. The floor file's note records the shape it was measured in — a GitHub runner with none of
+# this instance's local prerequisites, 26 skips. The EXEC MIRROR is a different shape: every local
+# prerequisite is present (it exists only on the machine the system runs on) but it is a flat file
+# copy, so it is not a git repository and has no `.github/`, and it skips 31 tests that a runner
+# does not skip at all. The result was a `verify.py` that was RED FROM THE MIRROR on every input
+# including a correct tree — and CLAUDE.md §1 makes the mirror run the verdict, so the instrument
+# that exists to catch cross-tree divergence was protecting nothing. A gate that is red whatever
+# you do gets ignored, then switched off.
+#
+# Raising `skipped_max` to 31 would have been the wrong repair: CI runs where 26 is right, and the
+# extra five would have let a runner skip five more checks in silence. So each shape carries its
+# own agreed number, measured in the shape it bounds — the measuring window equal to the draining
+# window, which is the house rule this file already applies everywhere else.
+#
+# ONE RULE, NOT A SECOND TABLE. The mirror's key for any ceiling is its base key with this prefix,
+# derived here and consumed by both `_ceiling_problems` and `--update-floor`, so the check and the
+# writer cannot name different keys and a ceiling added later gets its shape variant for free. A
+# mirror key that is unset (or misspelt) simply falls back to the base agreement — the STRICT
+# direction, which fails toward a loud red rather than toward silence.
+MIRROR_CEILING_PREFIX = "mirror_"
+
+
+def mirror_key(key: str) -> str:
+    """The floor-file key holding the EXEC MIRROR's agreed maximum for `key`."""
+    return f"{MIRROR_CEILING_PREFIX}{key}"
+
+
+def ceiling_limit(floor: dict, key: str, *, mirror: bool) -> tuple[int | None, str]:
+    """The limit in force for `key` in THIS tree, and the floor key that set it.
+
+    Returns the key as well as the number because every failure message here has to name the
+    integer to raise, and on the mirror that is a DIFFERENT key. "raise `skipped_max`" read from a
+    mirror run sends the reader to edit the number CI depends on — the gate telling you to break
+    the other environment to clear this one.
+
+    The mirror value REPLACES the base rather than adding to it. On the mirror the local
+    prerequisites are all present, so the runner-shape absences the base number pays for do not
+    occur; adding the two would buy 26 units of headroom that nothing there can legitimately spend.
+    """
+    if mirror:
+        mk = mirror_key(key)
+        limit = floor.get(mk)
+        if limit is not None:
+            return int(limit), mk
+    base = floor.get(key)
+    return (None if base is None else int(base)), key
+
+
 def mypy_exempt_modules() -> list[str] | None:
     """Modules on `[[tool.mypy.overrides]] ignore_errors` — the ratchet's blocking quantity.
 
@@ -692,8 +760,57 @@ def _exempt_ceiling_input(exempt: list[str] | None) -> int | None:
     return len(exempt) if exempt is not None else None
 
 
-def _ceiling_problems(floor: dict, actual: dict) -> list[str]:
-    """Is anything skipping MORE than the agreed maximum? Pure, for the same reason.
+def _preserved_ceilings(floor: dict) -> dict:
+    """Every agreed ceiling `--update-floor` must carry forward — BOTH shapes'. Pure, so the
+    preservation is asserted rather than trusted.
+
+    THE RECONCILE JOB REWRITES THIS FILE ON EVERY PUSH TO MAIN, from a checkout, where the mirror
+    ceilings are never consulted. A writer that kept only the keys the run happened to READ would
+    therefore delete them — silently re-latching the mirror shut, with the deletion arriving as a
+    bot commit nobody reads. It derives the key from `mirror_key`, the same rule the check uses; a
+    second list of key names would be a matching pair, and a drifted pair here is the deletion.
+    """
+    out: dict = {}
+    for key, _label in CEILINGS:
+        for k in (key, mirror_key(key)):
+            if floor.get(k) is not None:
+                out[k] = int(floor[k])
+    return out
+
+
+def _ceiling_report(floor: dict, actual: dict, *, mirror: bool) -> tuple[list[str], dict[str, str]]:
+    """The ceiling VERDICT and the `count/limit` RENDERING for every ceiling, from ONE shape
+    decision. Pure.
+
+    THEY ARE RETURNED TOGETHER ON PURPOSE. Two readers of one quantity that reach the shape
+    separately can disagree, and this file has already been bitten by exactly that: the mypy
+    ratchet's LINE said "NOT COUNTED" while its ENFORCEMENT cleared, because the renderer and the
+    checker coerced an uncountable value differently and the permissive one decided the exit code.
+    With the shape it would have been worse — a summary reading `31/31 max` above a verdict
+    computed against 26, which reads as a bug in the tool rather than in the tree. One call, one
+    decision, and the selftest can hold both halves against each other.
+    """
+    problems = _ceiling_problems(floor, actual, mirror=mirror)
+    rendered = {}
+    for key, _label in CEILINGS:
+        limit, in_force = ceiling_limit(floor, key, mirror=mirror)
+        observed = actual.get(key)
+        if limit is None:
+            rendered[key] = f"{observed} (no ceiling set)"
+            continue
+        # Name the KEY whenever it is not the base one. A bare "31/31 max" on the mirror looks
+        # like the number CI enforces and is not; the reader has to be able to see which
+        # agreement they are reading without going to look it up.
+        via = "" if in_force == key else f" [{in_force}]"
+        rendered[key] = f"{observed}/{limit} max{via}"
+    return problems, rendered
+
+
+def _ceiling_problems(floor: dict, actual: dict, *, mirror: bool = False) -> list[str]:
+    """Is anything skipping MORE than the agreed maximum FOR THIS TREE SHAPE? Pure, same reason.
+
+    `mirror` defaults to False so an unshaped call is bounded by the base (smaller) agreement —
+    the strict direction, and the one every existing caller and selftest already means.
 
     An UNSET ceiling means "nothing agreed yet", not "zero" — reading a missing key as 0 would
     condemn every machine that legitimately lacks a prerequisite.
@@ -706,14 +823,17 @@ def _ceiling_problems(floor: dict, actual: dict) -> list[str]:
     silence is the wrong answer.
     """
     problems = []
+    where = " (exec mirror)" if mirror else ""
     for key, label in CEILINGS:
-        limit = floor.get(key)
+        # The key that SET the limit, not the base key — on the mirror they differ, and every
+        # message below has to send the reader to the number that actually governs this run.
+        limit, in_force = ceiling_limit(floor, key, mirror=mirror)
         if limit is None:
             continue
         observed = actual.get(key, 0)
         if observed is None:
             problems.append(
-                f"CEILING UNCHECKABLE: {label} could not be counted, but `{key}` is set to "
+                f"CEILING UNCHECKABLE: {label} could not be counted, but `{in_force}` is set to "
                 f"{limit}. A bounded quantity that cannot be measured is a failure, not a pass — "
                 f"fix whatever made it uncountable, or remove the ceiling deliberately."
             )
@@ -723,9 +843,9 @@ def _ceiling_problems(floor: dict, actual: dict) -> list[str]:
                 # `CEILING`, not `SKIP CEILING`: the mypy exempt list is bounded by this same
                 # machinery and is not a skip, so the old wording sent a reader hunting for a skip
                 # that does not exist. The label already names WHICH population overflowed.
-                f"CEILING exceeded: {observed} {label} > agreed maximum {limit}. "
+                f"CEILING exceeded{where}: {observed} {label} > agreed maximum {limit}. "
                 f"This is bounded on purpose — either the new one is wrong, or raise "
-                f"`{key}` in .verify-floor.json deliberately and say why."
+                f"`{in_force}` in .verify-floor.json deliberately and say why."
             )
     return problems
 
@@ -779,12 +899,17 @@ def verify(
             "collection floor nor any ceiling can be checked. That is a failure, not an unset "
             "floor — fix the file rather than letting a run pass unmeasured."
         )
-    problems += _ceiling_problems(floor, actual)
+    # WHICH TREE IS THIS? Detected from the tree itself — never from `$CI` — because the exec
+    # mirror and a bare runner are two different deprived shapes and one number cannot bound both.
+    # See `ceiling_limit`. Computed once and threaded through the check and the rendering, so the
+    # verdict and the summary cannot disagree about which agreement was in force.
+    mirror_reason = exec_mirror_shape()
+    ceiling_problems, ceiling_rendered = _ceiling_report(floor, actual, mirror=bool(mirror_reason))
+    problems += ceiling_problems
 
     def _cap(key: str) -> str:
         """Both numbers, always in the same place: the count AND what bounds it."""
-        limit = floor.get(key)
-        return f"{actual[key]}" + (f"/{limit} max" if limit is not None else " (no ceiling set)")
+        return ceiling_rendered[key]
 
     # The floor reports its RELATIONSHIP to reality, not just its value. "floor 386" reads as
     # fine at a glance; "floor 386 — 1 BEHIND" cannot be misread, which is the same house rule
@@ -810,6 +935,13 @@ def _cap(key: str) -> str:
         )
     )
     lines = ["# verify.py", ""]
+    # ALWAYS printed, in both shapes. A line that only appears on the mirror tells a reader
+    # nothing on the machine where it is silent, and the whole point is that "which ceilings are
+    # in force" must never have to be inferred.
+    lines.append(
+        f"  tree:       {'EXEC MIRROR — mirror_* ceilings apply' if mirror_reason else 'checkout'}"
+        + (f"\n              {mirror_reason}" if mirror_reason else "")
+    )
     lines.append(
         f"  pytest:     {py['passed']} passed, {py['failed']} failed, "
         f"{_cap('skipped_max')} skipped "
@@ -893,9 +1025,7 @@ def _cap(key: str) -> str:
         # Recording bare `passed` from a skipping machine would RATCHET THE FLOOR DOWN by exactly
         # the amount that was skipped — the floor following the leak instead of catching it.
         blob = {"collected": py["collected"], "passed": py["passed"] + py["skipped"]}
-        for key, _label in CEILINGS:
-            if floor.get(key) is not None:
-                blob[key] = int(floor[key])
+        blob.update(_preserved_ceilings(floor))
         blob["note"] = _appended_note(
             str(floor.get("note", "")),
             py["collected"],
@@ -1019,6 +1149,93 @@ def _selftest() -> None:
         == []
     )
 
+    # ---- TWO SHAPES, TWO AGREEMENTS (2026-08-29) ----------------------------------------------
+    # `skipped_max` bounded a GitHub runner (26, per the floor note) and was ALSO applied to the
+    # exec mirror, which skips a different 31 — so a mirror run was red on every input including a
+    # correct tree, and CLAUDE.md §1 makes that run the verdict. The whole repair is that the
+    # limit in force depends on the detected shape, so that is what is asserted, in both
+    # directions and on the permissive side especially.
+    _two = {"skipped_max": 26, "mirror_skipped_max": 31}
+    _s31 = {**{k: 0 for k, _ in CEILINGS}, "skipped_max": 31}
+    # The bug, reproduced: on a CHECKOUT 31 skips is still a breach and must stay one.
+    _as_checkout = _ceiling_problems(_two, _s31, mirror=False)
+    assert len(_as_checkout) == 1 and "CEILING exceeded" in _as_checkout[0], _as_checkout
+    assert "`skipped_max`" in _as_checkout[0], "a checkout must be sent to the base key"
+    # The fix: the same 31 on the MIRROR is exactly at its own agreed maximum, and passes. This is
+    # the assertion that would have been red for as long as the defect existed — the drained
+    # state, which the house rule says must be reachable by some real input. It is: this is the
+    # live mirror's measured count.
+    assert _ceiling_problems(_two, _s31, mirror=True) == [], "the mirror must be able to go GREEN"
+    # And the mirror ceiling is still a ceiling: one over it fails, naming the MIRROR key. Sending
+    # a mirror reader to raise `skipped_max` would tell them to loosen the number CI depends on.
+    _over_m = _ceiling_problems(_two, {**_s31, "skipped_max": 32}, mirror=True)
+    assert len(_over_m) == 1 and "`mirror_skipped_max`" in _over_m[0], _over_m
+    assert "exec mirror" in _over_m[0], "the message must say which shape it is bounding"
+    # A ceiling with NO mirror variant falls back to the base agreement — the strict direction, so
+    # a misspelt or forgotten key fails loudly instead of going unbounded.
+    _fallback = _ceiling_problems(
+        {"selftest_skipped_max": 7},
+        {**{k: 0 for k, _ in CEILINGS}, "selftest_skipped_max": 8},
+        mirror=True,
+    )
+    assert len(_fallback) == 1 and "`selftest_skipped_max`" in _fallback[0], _fallback
+    # `mirror=False` is the default, and it must be the STRICT one: an unshaped call cannot help
+    # itself to the looser number.
+    assert _ceiling_problems(_two, _s31) == _as_checkout, "the default shape is checkout"
+    # The key rule itself, since both the check and `--update-floor` derive from it. A second
+    # table would be a matching pair, and a drifted pair here deletes the mirror ceiling.
+    assert mirror_key("skipped_max") == "mirror_skipped_max"
+    assert ceiling_limit(_two, "skipped_max", mirror=True) == (31, "mirror_skipped_max")
+    assert ceiling_limit(_two, "skipped_max", mirror=False) == (26, "skipped_max")
+    assert ceiling_limit({"skipped_max": 26}, "skipped_max", mirror=True) == (26, "skipped_max")
+    assert ceiling_limit({}, "skipped_max", mirror=True) == (None, "skipped_max")
+    # THE REAL FLOOR FILE must actually carry the mirror agreement, or every assertion above is
+    # about a shape that nothing configures — the defect would still be live with a green suite.
+    _real_floor = load_floor()
+    assert _real_floor.get("mirror_skipped_max") is not None, (
+        ".verify-floor.json records no `mirror_skipped_max`; a mirror run is then bounded by the "
+        "runner's number again, which is red on every input"
+    )
+    # NOT ASSERTED: that the mirror ceiling is >= the base one. It is today (31 vs 26) and that is
+    # a coincidence, not an invariant — the two populations are DISJOINT, one runner-shape and one
+    # mirror-shape. Worse, the assertion would have latched: the drain named in the floor note is
+    # teaching orch-sync-mirror.sh to copy `.github/`, which drops 12 mirror skips and takes this
+    # number to 19, and a `>=` here would then refuse the drain it exists to encourage. A gate
+    # whose clear path is blocked by the thing it measures is the defect this whole file is about.
+    # And `--update-floor` must PRESERVE it. The reconcile job rewrites this file on every push to
+    # main, from a checkout, where the mirror keys are never consulted — a writer that kept only
+    # the keys it used would delete the ceiling and silently re-latch the mirror shut.
+    assert _preserved_ceilings({"skipped_max": 26, "mirror_skipped_max": 31, "note": "n"}) == {
+        "skipped_max": 26,
+        "mirror_skipped_max": 31,
+    }, "--update-floor must carry BOTH shapes' agreements forward, and nothing else"
+    assert "mirror_skipped_max" in _preserved_ceilings(
+        _real_floor
+    ), "--update-floor would drop the real mirror ceiling on the next push to main"
+    # THE VERDICT AND THE SUMMARY MUST AGREE, which is why they come from one call. Held against
+    # each other here, in both shapes: a run that renders `31/31 max` above a verdict computed
+    # against 26 reads as a broken tool, and the reader stops believing either number.
+    _probs_m, _rend_m = _ceiling_report(_two, _s31, mirror=True)
+    assert _probs_m == [] and _rend_m["skipped_max"] == "31/31 max [mirror_skipped_max]", _rend_m
+    _probs_c, _rend_c = _ceiling_report(_two, _s31, mirror=False)
+    assert len(_probs_c) == 1 and _rend_c["skipped_max"] == "31/26 max", _rend_c
+    # An unset ceiling still says so rather than rendering a bare count.
+    assert "no ceiling set" in _ceiling_report({}, _s31, mirror=True)[1]["skipped_max"]
+    # THE ONE WIRING PIN, because everything above is pure and `verify()` is not cheap to call:
+    # dropping the keyword reverts the defect exactly (the mirror bounded by the runner's number)
+    # while every assertion above stays green — this repo's founding failure, code that exists and
+    # is not invoked. Split so the needle cannot match its own line, and the fragment is the
+    # smallest thing that would be ABSENT if the wiring were removed: no newline, no indentation,
+    # no argument list a formatter owns.
+    _src = pathlib.Path(__file__).read_text(encoding="utf-8")
+    assert ("mirror=bool(" + "mirror_reason)") in _src, (
+        "verify() no longer feeds the DETECTED shape into the ceilings, so the exec mirror is "
+        "bounded by the runner's `skipped_max` again and is red on every input. This is the one "
+        "pin here: everything else about the shape rule is pure and asserted above, but verify() "
+        "itself is too expensive to call from a selftest, and a hardcoded False would revert the "
+        "defect exactly while every assertion above stayed green."
+    )
+
     # ---- the floor counts what RAN OR WAS NAMED, and the ceiling stops that being a loophole --
     # 330 -> 300 passed with 30 named skips is fine; 300 passed with 0 skips is a test that
     # vanished. Both directions, because only having one is how a floor becomes decoration.
@@ -1320,7 +1537,8 @@ def _selftest() -> None:
         "--floor-may-lag forgives a lagging floor on a PR but never a collection drop, "
         "--reconcile-floor forgives ONLY a drift it "
         "healed and never an unwritten one, absent-module line is silent when clean and is "
-        "never counted as a skip, mypy ratchet prints both numbers and its ceiling can fail)"
+        "never counted as a skip, mypy ratchet prints both numbers and its ceiling can fail, "
+        "two tree shapes carry two agreed ceilings and the exec mirror can go GREEN)"
     )