[noahgift]

Manual batch (operator ruling, checkin-ruling-1615 item 1): six ready branches cherry-picked onto origin/main in one PR. make gate ran green on lambda (rc 0, log /mnt/nvme-raid0/targets/batch1-gate3.log).

Tickets folded (each source PR keeps its existing seats; all test-only or script-only, no gate or workflow file touched):

• test(simulate): cover validate_verification_tests (EDD-07) #5735 test(simulate): validate_verification_tests (EDD-07)
• test: fold train loader try_load_apr coverage #5592 test: train loader try_load_apr coverage
• test: fold #5538 DataCard::paint coverage #5586 test: DataCard::paint coverage
• test(apr-cli): fold run print_run_output coverage (FALSIFY-OPS-013) #5657 test(apr-cli): run print_run_output coverage (FALSIFY-OPS-013)
• test(registry): fold DownloadProgress human accessors coverage #5672 test(registry): DownloadProgress human accessors coverage
• contracts.nt + witness are generated shared files: every contract PR conflicts (regen-at-merge vs fragment) #5762 (work/5762 @9ff4016841): scripts/contracts_resolve.sh + proposal doc

Only six qualified, not seven. #5752 entered the merge queue solo before this PR and was dropped from the fold. #5589, #5583 and #5527 were excluded (queued solo, or touching shared files). Derived files (contracts.nt, README count) were regenerated, not hand-merged. The source PRs become redundant on merge; closing them is the operator's.

keep-open: none
ont-delta: none — test, contract-data and doc-comment changes only; no ontology claim, level or cell moves (README counts regenerated by readme_sync)

🤖 Generated with Claude Code

[noahgift]

seat 1, aprender-pace @281d6cf2ea22105d56c4c1b4be29f63c1d13166e

Read every constituent diff (16 files, merge-base diff). Live head confirmed; 0 aprender-pace trailers in the PR's commits.

• #5735 simulate EDD-07 tests, #5592 try_load_apr tests, #5586 DataCard::paint tests, #5657 print_run_output + FALSIFY-OPS-013, #5672 DownloadProgress human accessors: test-only, each asserts exact values (no vacuous asserts); each new contract yaml matches its test. README count 1985 -> 1989 = the 4 new contract yamls.
• #5762: scripts/contracts_resolve.sh (new, refuses non-generated conflicts, takes no side, the gate stays the judge) + proposal doc. Not wired into any gate, workflow or Makefile; no gate content change.
• The batch commit "drop #5752 fold" leaves no #5752 content in the diff.

Not verified: I did not run any test or the gate; CI checks were still pending at review time; contracts.nt (+20 lines, 4 contracts x 5 triples) not regenerated by me, and the witness file is unchanged (plausible for kind=pattern contracts, not re-derived).

[noahgift]

seat 2, aprender-triage-worker @281d6cf2ea22105d56c4c1b4be29f63c1d13166e APPROVE

Own-trailer check: 0 of 13 commits carry a triage-worker trailer; I authored none of it.

Read (every constituent diff, 16 files, +564/-2):

• apr-cli run_tests_accel_reconcile.rs (+28): test sits inside mod tests of run.rs (run.rs:705 include!(run_07.rs) -> run_07.rs opens #[cfg(test)] mod tests and includes the file), so FALSIFY-OPS-013's path commands::run::tests::print_run_output_serves_every_mode_and_derives_the_rate resolves.
• contracts/apr-cli-operations-v1.yaml (+FALSIFY-OPS-013), 4 new contract YAMLs, contracts.nt (+20), README count 1985->1989: consistent with each other.
• aprender-registry cache.rs (+31), aprender-simulate edd/validation.rs (+50), aprender-present-widgets data_card.rs (+106), aprender-train config/train/loader/tests.rs (+75): test-only, assertions are non-vacuous.
• docs/findings/aprender-w5404.jsonl, w5382.jsonl (+1 each); docs/proposals/contracts-nt-regen-at-merge.md (+39, read in full); scripts/contracts_resolve.sh (+42): dev helper, refuses unless every conflict is a generated path, not on a CI path.

Not run: no local build or test run. The witness JSON is not in this PR's file set; the contracts gate judges it.
Notes: PR is BEHIND and its 6 checks had no conclusion when I looked, so this seat is not a green-CI claim. Size is +564 across 6 folds, which is over plan step 5 (<=200 lines, one ticket per PR); that is the cop's call, not a code defect.

[noahgift]

seat 1, aprender-pace @92e5602b5800a047c1fda177d04acc6d9fef6b3a

Re-seat. Delta since 281d6cf is one commit, one file: contracts/witness/811c7818...d850.json, 3+/3-. Live head confirmed; 0 aprender-pace trailers in the PR.

• census_id_set_sha256 changed (expected: 4 new contract ids); relations_sha256 and filename unchanged; result/checkable_n/pc_reasoner unchanged; reasoner_git_sha = the previous head 281d6cf; cpu_ms 2 -> 1.
• Everything else is byte-identical to what I read at 281d6cf (my earlier seat's findings stand: test-only folds + contracts_resolve.sh, no gate content).

Not verified: I did not re-run pv-sat or the gate to reproduce the new census hash, tests not run, CI results not read.

[noahgift]

seat 2, aprender-triage-worker @92e5602b5800a047c1fda177d04acc6d9fef6b3a APPROVE

Own-trailer check: 0 commits carry a triage-worker trailer.
Delta from my seat at 281d6cf: one file, the witness JSON (3+/3-): census_id_set_sha256 and reasoner_git_sha updated to the prior head, cpu_ms 2->1; relations_sha256 and result unchanged. That closes the witness-not-in-file-set note from my first seat. The other 15 files are as I read them at 281d6cf.
Not run: no local build or test; the contracts gate judges the witness. PR is BEHIND. Size over plan step 5 stands as the cop's call.

[noahgift]

seat 2, aprender-triage-worker @7b8f90f61d8db601fbdadd09207114b6b1ffbe6d APPROVE (re-seat; prior seat @92e5602b58)

Own-trailer check: 0 of 15 commits carry a triage-worker trailer.
Delta from my last seat, restricted to the PR's current file set (17 files): git diff 92e5602b58 7b8f90f61d -- <those files> is empty, so the head differs only by merges of origin/main (81e7bc0 #5745 and b0ec584 #5732 landed since). Everything I read at 92e5602 stands. Not run; PR BLOCKED at read, CI not judged.

[noahgift]

seat 1, aprender-pace @7b8f90f61d8db601fbdadd09207114b6b1ffbe6d

Re-seat. 0 aprender-pace trailers. Delta vs my seat @92e5602b58 = one merge of origin/main (81e7bc0; parent 2) and nothing else: the per-side PR diff (generated files excluded) is empty, and the witness json is byte-identical to 92e5602's. Net diff vs the new merge-base (81e7bc0) is the same 17 files as before, with no shapes.ttl change (the 68-line shapes.ttl delta vs 92e5 is main's own earlier change brought in by the merge, not this PR's).

Not verified: tests/CI not run or read; the witness census hash not reproduced with pv-sat.

[noahgift]

seat 2, aprender-triage-worker @dbe7ec1317f724c6f828ae52435dfdce6ff12d27 APPROVE (re-seat; prior seat @7b8f90f61d)

Own-trailer check: 0 commits carry a triage-worker trailer.
Delta from my last seat in the PR's file set: one file, the witness JSON 811c7818...json (+1/-3). It drops reasoner_git_sha and cpu_ms, the per-run fields. That matches the shape of the witness on origin/main now (no per-run fields; census_id_set_sha256 differs only because this PR adds contracts, and it is the same 59167d32... value as at my earlier seat). The other files are unchanged from what I read. I did not regenerate the graph, so the hash correctness is judged by the contracts gate, not by me. Not run; PR BLOCKED at read, CI not judged.

[noahgift]

seat 1, aprender-pace @dbe7ec1317f724c6f828ae52435dfdce6ff12d27

Re-seat. 0 aprender-pace trailers. Delta vs my seat @7b8f90f61d = one merge of origin/main (bf957b2, which brings #5752) plus one regen commit. The PR's own diff (generated files excluded) is identical to 7b8f90f's.
The only contracts change is the witness json (811c7818...): reasoner_git_sha and cpu_ms are dropped. That is what #5752 does (it stops writing those volatile fields; witness.rs gained skip_serializing on them), so the regenerated file matches main's new format. The witness filename (relations hash) is unchanged.

Not verified: tests/CI not run or read; the witness was not re-derived with pv-sat.

[noahgift]

seat 1, aprender-pace @98fae558e33cec374c2d46b9f80ea62db594616a

Re-seat. 0 aprender-pace trailers. Delta vs my seat @b99cef1efe = two commits (bf2a725, 98fae55), net 2 lines in one file: crates/aprender-simulate/src/cli/args.rs, the doc-comments that become the sim repo-flow and course-flow help lines. "; --json prints the sim block" is dropped from both, so neither line is long enough to wrap under clap's wrap_help. No other file changed, and no other place in the tree cites the dropped text. The quorum receipt (quorum/quorum-helpwrap-98fae558e3...json: 3 opus lanes, 3 APPROVE) is in the cop inbox; I read it and did not redo it. Its notes are open caveats, not blockers: the help-line parser is fragile (follow-up ticket) and the test depends on COLUMNS.

Not verified: cli_surface_gate / tests / CI not run or read; I did not reproduce the original red.

[noahgift]

seat 2, aprender-triage-worker @98fae558e33cec374c2d46b9f80ea62db594616a APPROVE (re-seat)

0 triage-worker trailers on the PR. Delta from my last seat (dbe7ec1) restricted to the PR's file set: the only changes since are b99cef1 (contract pass_criteria count), bf2a725 and 98fae55. The last two touch only crates/aprender-simulate/src/cli/args.rs: two doc-comment lines (RepoFlow, CourseFlow), each losing its "; --json prints the sim block" tail, so clap's wrap_help no longer splits the line. No code path changes.

Read: the compare diff b99cef1...98fae55 (2 additions, 2 deletions) and the quorum receipt for this head (3/3 APPROVE). Not run: no build or tests, so I did not check that cli_surface_gate turns green. Noted from the receipt, non-blocking: the help-line parser is fragile and a test depends on COLUMNS (follow-up ticket).

[noahgift]

seat 2, aprender-triage-worker @1a160b75f5e74c9d2237a5908127eb088fed6d36 APPROVE (re-seat)

0 triage-worker trailers on the PR's commits. Delta from my last seat (98fae55), restricted to the PR's own commits: 75f1a1f (ledger rows, counts), 5645260 (merge of main) and 1a160b7 (contracts.nt regen). Everything else in the compare is main's content arriving through the merge.

Question asked: does 75f1a1f change only ledger data, not what the gate accepts? Yes. It touches three files and none is a script, workflow or check: docs/audits/surface_audit.csv (+4 rows), contracts/apr-dogfood-coverage-v1.yaml (counts, not the rules) and one number in .claude/skills/apr-dogfood/SKILL.md. The four rows are apr sim repo-flow, apr sim course-flow and the two aprender-simulate equivalents, hardware x86_64-linux, gate NONE, evidence args.rs:169/181. I recounted the CSV at the head: 911 rows, 390 apr, 15 aprender-simulate, 23 cluster simulation, which match the yaml's 911 / 390 / 15 / 23 and the 144/911 ratio. covered stays 144 and the UNKNOWN ceiling stays 427 (the new rows have known hardware, so the ceiling cannot be exceeded by them). scripts/tree_reader_tests.txt in the compare comes from main, not this PR.

Not run: no build, no check_dogfood_coverage.sh / G2.1, no pv validate. I did not verify the claim that the four rows were run on x86_64; the sim's own unit tests exercise both flows, but the PR carries no run receipt. That is the same question that failed the earlier 5590-b round, so I flag it as non-blocking here: the rows add no gated coverage.

[noahgift]

seat 1, aprender-pace @1a160b75f5e74c9d2237a5908127eb088fed6d36

Re-seat delta vs my seat at 98fae55 (75f1a1f + main merge + contracts.nt regen). Checked: the PR's own change, with generated files and README excluded, differs from 98fae55 only by ledger data.

• docs/audits/surface_audit.csv: +4 rows (apr sim repo-flow/course-flow, aprender-simulate repo-flow/course-flow), x86_64-linux, in_dogfood_skill=no, cluster 11, evidence args.rs:169/:181.
• Counts re-derived from the CSV at this head: rows 911, apr 390, aprender-simulate 15, simulation cluster 23, in_dogfood=yes 144. They match the bumps in apr-dogfood-coverage-v1.yaml (907->911, 388->390, 13->15, 19->23, q5_6 719->723, ratio 0.1581 = 144/911) and SKILL.md. Covered stays 144, so the 4 rows are uncovered.
• No script, ci/ or workflow file changed in this delta. The only scripts/ file in the PR is contracts_resolve.sh, already seated and wired into no gate. What G2.1 accepts is unchanged; the rows add ledger entries so it sees the args.rs edit.
  Not verified: that G2.1 now passes (no CI run), and the contracts.nt regen (not re-derived).

[noahgift]

seat 1, aprender-pace @3f5488c03692a8936487c44dfcc03199e40efab2

Re-seat delta vs my seat at 1a160b7: merge origin/main + regen only. Checked:

• The PR's own change (generated files and README excluded) is identical to 1a160b7: the diff of the two is empty.
• Witness: NOT byte-identical to main's. Filename (relations_sha256) is main's; the only differing line is census_id_set_sha256, which is what 4 added contracts should move. It tracks the README count 1987 -> 1991 and the contracts.nt additions (+32 lines only, no deletions).
• README resync: the two CONTRACT_COUNT lines only, 1987 -> 1991.
  Not verified: I did not re-run pv-sat/pv extract, so the new census_id_set_sha256 and the nt lines are not re-derived; no CI/compile.

[noahgift]

seat 2, aprender-triage-worker @3f5488c036 APPROVE (re-seat by delta from my seat at 1a160b7). Read: git log/diff 1a160b7..3f5488c = one merge of main (b08ba56, #5773) with regenerated derived files. Checked vs origin/main b08ba56: README count 1987->1991 = +4 contract yamls added on the PR; witness differs only in census_id_set_sha256 (same file name, relations hash unchanged); contracts.nt +32/-0 lines, additive; PR file set otherwise unchanged. 0 own trailers. Not run: contracts_gate.sh, CI jobs.

[noahgift]

seat 1, aprender-pace @d8f12f4d507a9a81d1d41ed34591b0bc16c63470

Re-seat delta vs my seat at 3f5488c: a merge of main (9646df7, #5583) + regen. Checked:

• The PR's own change (generated files, README and w5382.jsonl excluded) is identical to 3f5488c: the diff of the two is empty.
• docs/findings/w5382.jsonl, the hand union: merge-base 2 lines, ours (3f5488c) 3, main (9646df7) 3, head 4. Every line of ours and of main is in head, none is duplicated (4 unique of 4), the added line from each side is distinct, every line parses as JSON, no conflict markers, file ends with a newline. So base 2 + one from each side = 4, nothing lost.
• Witness: same filename as main's; only census_id_set_sha256 differs from main (the PR's contracts). README: only the two CONTRACT_COUNT lines, 1989 -> 1993 (the same +4 this PR has carried). contracts.nt vs main: additions only, no deleted lines.
  Not verified: I did not re-run pv-sat/pv extract, so the new census_id_set_sha256 and the nt lines are not re-derived; no CI/compile. 0 aprender-pace trailers.

[noahgift]

seat 2, aprender-triage-worker @d8f12f4d50 APPROVE (re-seat by delta from my seat at 3f5488c). Read: log 3f5488c..d8f12f4 = one merge of main (9646df7, #5583). PR file set shrank by exactly the 4 files #5583 now carries on main (blis elementwise + its contract, pv-query-diff-info + its tests). docs/findings/w5382.jsonl hand-union checked: main 3 lines, merged file 4 = main's 3 + the PR's 1 (FND-20261018-registry-progress-human-untested); 0 main lines missing, 0 old-head lines missing, 0 duplicates. aprender-w5404.jsonl and surface_audit.csv have no duplicate lines. Derived files vs main: README 1989->1993 = the 4 contract yamls the PR adds; witness differs only in census_id_set_sha256; contracts.nt +32/-0 additive. 0 own trailers. Not run: contracts_gate.sh, CI.