[noahgift]

Batch 4 (SIM-4 only)

SIM-4 capacity-bound check for the PR-flow prediction (dc0d875), its contract regen (8123d29) and the runs_per_pr mutant kill (92c835e), rebased on post-#5773 main with derived files regenerated (contracts_gate regen, readme_sync). 3 commits, ~129 lines outside derived files, touches no gate or workflow file.

5590-b is NOT in this batch: its bin edit forces a ledger touch (G2.1) and flipping the hardware column fails G5.1.

Local gate (lambda, alone): validate_contracts 10/10, tree-reader check PASS, dogfood coverage gate PASS, cargo test -p aprender-simulate --lib pr_flow 19/19, cargo check --workspace --all-targets finished. Workspace nextest run: 9843 passed, 1 flaky (apr-cli eval verbose_stderr test timed out at 120s under load, passed on retry), then a cargo-memcap KILL on apr-cli golden_output (same test dies on main) stopped the run: 6445 tests not run = not_measured. CI's full gate decides.

Quorum: quorum/quorum-batch4-sim4-only.json (opus lanes; follow-up: the SIM-4 contract text says every batch is checked but sim4_breaches has no non-test caller yet).

keep-open: no
ont-delta: reason README CONTRACT_COUNT 1993 -> 1994 from the new contracts/pr-flow-sim4-bounds-v1.yaml, regenerated by readme_sync; no ontology claim changes

🤖 Generated with Claude Code

[github-actions]

§13.11 rung 1 — quorum shadow verdict

S13-SHADOW pr=5774 head=25bcc08851cd54060203a5586e87c31ec5bcd7a2 verdict=REFUSE class=Q1 arm_rc=1

Shadow mode: this records a verdict and merges nothing. A refusal
to arm is not a block (§13 adds zero rows to §7) — the pull request is
exactly as green as it was.

[noahgift]

seat 2, aprender-triage-worker @99dd1f1c1f APPROVE. Read: full diff vs origin/main (5 files, +137/-3): sim4_breaches in pr_flow.rs (three inequalities, 51 = 85% of 60, cap carries the 0.5% SIM-3 tolerance, names the breached ones), 3 tests (reference designs D/G/E clean and A flags mq_passes; each bound named alone via cap/r_pr/r_mq overrides plus a 1.5x all-three case; runs_per_pr scaling kills the rework mutant), contract pr-flow-sim4-bounds-v1.yaml (3 equations, 3 falsifiers each naming a real test), regen commit (README 1987->1988 = +1 yaml, witness census hash only, contracts.nt +7 additive). No surface_audit.csv rows, so none of the 5590-b unearned hardware rows. 0 own trailers. Not run: cargo tests, contracts_gate.sh, CI.

[noahgift]

seat 2, aprender-pace @99dd1f1c1f938491faf51b2619027c717e6f587d

Read the whole PR (5 files, 137 insertions): sim4_breaches + RUNNER_MIN_PER_H in scenarios/pr_flow.rs, three tests, the new contract pr-flow-sim4-bounds-v1.yaml, regen.

• Logic matches the contract's three equations (merged/h <= cap/W with the 0.5% Little tolerance, PR runner-min vs R_pr51, MQ passes vs R_mq51/mq_runner_min). Fields and types used all exist (usize casts to f64, PrFlowResult: Clone).
• Contract: 3 FALSIFY ids each name a real test, and the tests exist under those exact names.
• Ran cargo test -p aprender-simulate --lib sim4 at this head: 3 passed, 0 failed (so D/G/E breach nothing and A flags exactly mq_passes).
• Generated files: witness differs from main by the census_id_set_sha256 line only (one new contract); README changes only the two count lines; contracts.nt is additions only. 0 aprender-pace trailers in the PR.
  Not verified: clippy/fmt, the full CI, and re-derivation of the witness sha / contracts.nt / README count via pv. sim4_breaches has no caller yet, so the "andon" the doc comment mentions is not wired; this PR adds the check only.

[noahgift]

seat 2, aprender-pace @3d7b74ae9e2a5e5771ac718214492f4e23b13144

Re-seat delta vs my seat at 99dd1f1: a merge of main (9646df7) + regen. The PR's own change (generated files and README excluded) is identical to 99dd1f1: the diff of the two is empty, so my earlier read of sim4_breaches, the contract and the tests stands.

• Witness: main's filename; only census_id_set_sha256 differs from main (the one new contract). README: only the two CONTRACT_COUNT lines, 1989 -> 1990. contracts.nt vs main: additions only.
  Not verified: I did not re-run the sim4 tests on this head (they passed at 99dd1f1), did not re-derive the witness sha / nt / README via pv, and did not check CI. 0 aprender-pace trailers.

[noahgift]

seat 2, aprender-triage-worker @3d7b74ae9e APPROVE (re-seat by delta from my seat at 99dd1f1). Read: log 99dd1f1..3d7b74a = one merge of main (9646df7, #5583); the old->new code delta is only #5583's files (blis elementwise, query_enrich_tests, w5382.jsonl). PR vs main is still the same 5 files, +137/-3. Derived: README 1989->1990 = +1 contract (pr-flow-sim4-bounds); witness differs only in census_id_set_sha256; contracts.nt +7/-0 additive. 0 own trailers. Not run: cargo tests, contracts_gate.sh, CI.

[noahgift]

seat 2, aprender-pace @25bcc08851cd54060203a5586e87c31ec5bcd7a2

Re-seat after the main merge + regen. The PR's own change (excluding contracts/witness, contracts.nt, README) is line-for-line identical to my seat at 3d7b74a: delta is empty. 0 aprender-pace trailers.
Not verified: tests, CI, clippy, pv re-derivation of the generated files at this head.

[noahgift]

seat 2, aprender-triage-worker @25bcc08851 APPROVE (re-seat by delta from my seat at 3d7b74a). Read: log 3d7b74a..25bcc08 = one merge of main (d5dedf1, #5765); old->new code delta is only #5765's files. PR vs main is unchanged: the same 5 files, +137/-3. Derived: README 1993->1994 = +1 contract (pr-flow-sim4-bounds); witness differs only in census_id_set_sha256 (no volatile fields); contracts.nt +7/-0 additive; shapes.ttl untouched. 0 own trailers. Not run: cargo tests, contracts_gate.sh, CI.