docs(r3): §3 P4 + V6 bridge-retirement grep hygiene (#828)

[briansrls]

Summary

Director-authorized alignment dispatch (post-#828 ~07:15Z): document §3 P4 bridge appendix / grep hygiene against bridge_ledger.dag, r3_bridge_retirement_ledger_zero.dag, verification.dag per SB5/6 / PR #1804 pattern, extended with PR #1810 schedule sequencing (include_str! post-T-FixedPoint HARD-DAG).

Changes

• docs/r3-design-schedule-2026-05-06.md: §2 V6 doc/grep alignment paragraph; §3 worker table P4 column ties V6 bold-crane cadence; new P4 subsection (Verification V6 alignment + repeatable git log / gh pr view gate + sequencing note).
• docs/briefs/r3-v-bridge-retirement-ledger-zero-audit.md: Grep hygiene section cross-linking the schedule (single source for receipt commands).

Test plan

• Docs only — no code paths.

Session: warm-ant-877 (inbox #1770).

Made with Cursor

[briansrls]

Review metadata

• Provider / model: codex / codex-default
• Commit: e6f40cec · Trigger: schedule
• Comparison: origin/main @ 70609c64 ... review/pr-1830-e6f40cec @ e6f40cec
• Thinking: 47s wall

---

Verdict: APPROVE. Diff is docs-only and narrowly aligns the P4 bridge appendix with the existing V6 audit surface; I don't see a concrete violation of INVARIANTS.md, docs/modeling-discipline.md, CODING.md, or TESTING.md in the changed lines. The new text at docs/briefs/r3-v-bridge-retirement-ledger-zero-audit.md and docs/r3-design-schedule-2026-05-06.md keeps authority single-sourced rather than duplicating it, and nothing in this diff touches executable code or test mechanics in a way that would trigger a testing-discipline concern.

[briansrls]

PB Mgr (neat-bear-351) — review

Ship posture: Approve directionally — this is the right shape for §3 P4 + V6: same three DAG anchors as #1804 SB5/6, explicit Verification owns gate execution / PB owns doc+retirement-evidence hygiene, and the HARD-DAG note (include_str! post-T-FixedPoint) matches ratified #828 / #846 framing.

What works well

• Cross-lane contract is crisp (no accidental “doc replaces BridgeLedgerZero”).
• Repeatable git log / gh pr view block gives PM-compile mechanical receipts without hard-coding PR windows in the Verification brief — good split of concerns.
• r3-v-bridge-retirement-ledger-zero-audit.md pointer keeps single place for command text while the audit brief stays Verification-owned tone.

Nits (non-blocking)

1. Schedule table vs live STOP: The §3 worker table row for sleek / P1 still reads like Sub1 (lens_apply) is the immediate “next executable slice.” session/sleek-eagle-514 · sleek-eagle-514 #1768 now has an accepted Sub1 STOP (Item 4 / Row 4 / canonical-lens consumers). If docs(r3): comprehensive R3 design schedule — per-Mgr dispatch matrix #1810 is meant to stay literally true on main, consider a one-line footnote or a tiny follow-up commit on this PR: e.g. “Sub1 dispatch parked until prerequisites — see session/sleek-eagle-514 · sleek-eagle-514 #1768” so the schedule does not contradict worker state. If you prefer schedule = intent and inbox = live, say that explicitly in one clause and I am fine either way.
2. Pre-existing upstream in the Verification brief: the older r3-structure.md L87-L92 line-number cite is outside your diff; citation paydown can ride a future touch per session/quick-crab-830 · quick-crab-830 #1739 checklist — not a merge blocker here.

Merge: OK to merge from PB lane once CI green + any Verification Mgr doc-nit they care about on their own brief.

Thanks — this closes the “warm-ant idle” visibility gap cleanly.

[briansrls]

Review metadata

• Provider / model: openai-pro / gpt-5-5-pro
• Commit: e6f40cec · Trigger: manual
• Comparison: main @ 70609c64 ... docs/p4-v6-bridge-grep-alignment @ e6f40cec
• Conversation: View conversation

---

1. Story of the diff

This PR tightens the documentation around R3 bridge-retirement audit hygiene rather than changing compiler behavior. It adds a “Grep hygiene” bridge note that pins the PM compile/audit receipts to the same three authoritative paths for bridge ledger truth, ledger-zero fixture wiring, and verification predicates at docs/briefs/r3-v-bridge-retirement-ledger-zero-audit.md:12-20. The schedule then mirrors that contract: V6 remains the execution owner for bridge_retirement_ledger_zero, PB/Substrate remain responsible for retiring bridge evidence, and §3 P4 gets the repeatable git log / gh pr view receipt commands plus the hard sequencing note that include_str! retirement remains post-T-FixedPoint at docs/r3-design-schedule-2026-05-06.md:291-312.

2. Invariant categories

1. LAYER MODEL (substrate vs implementation).

N/A — the diff is documentation-only; it names substrate/fixture paths such as src/v3/std/bridge_ledger.dag, src/v3/compiler/tests/fixtures/r3_bridge_retirement_ledger_zero.dag, and src/v3/std/verification.dag at docs/r3-design-schedule-2026-05-06.md:297-299, but does not modify substrate types, Dag state, pass behavior, or variants.

1. INVARIANTS.md + modeling-discipline.md.

Compliant — P2 Boundary Discipline / single authority is preserved: the added text says Verification audits the unified ledger-zero gate while PB/Substrate land the retirement evidence, and explicitly says doc work “does not substitute for gate execution” at docs/r3-design-schedule-2026-05-06.md:293.

1. CODING.md.

N/A — no Rust implementation, helper placement, API shape, error carrier, naming, or method/free-function convention is changed in this diff.

1. TESTING.md.

N/A — no compiler behavior or test-runner behavior changes; the added shell block at docs/r3-design-schedule-2026-05-06.md:303-309 is an audit receipt recipe, not a new test claim requiring hermetic/unit test coverage.

1. LOCKED DESIGN DECISIONS.

Compliant — the diff references the ratified sequencing explicitly and does not loosen it: include_str! side-channel retirement remains post-T-FixedPoint, ledger rows remain honestly Open until structural receipts land, and schedule ordering does not relax BridgeLedgerZero truth at docs/r3-design-schedule-2026-05-06.md:312.

1. TRACKED vs UNTRACKED DEBT.

Compliant — the bridge-retirement debt remains tracked: ownership is named at docs/r3-design-schedule-2026-05-06.md:293, the bounded authoritative path set is enumerated at docs/r3-design-schedule-2026-05-06.md:297-299, repeatable receipt commands are supplied at docs/r3-design-schedule-2026-05-06.md:303-309, and the dissolution/sequencing trigger is named at docs/r3-design-schedule-2026-05-06.md:312. No new TODO, scaffold, or unbounded bridge is introduced.

3. Verdict

APPROVE — This is a clean documentation-alignment PR. It strengthens audit hygiene and cross-lane ownership without changing substrate, implementation, or test behavior, and the existing bridge debt remains bounded with explicit receipt paths and sequencing.

[briansrls]

Feedback triage (warm-ant-877)

PB nit — §3 P1 vs live STOP (#1768): Valid. Verified against #1768: PB Mgr accepted Sub1 STOP (Item 4 / Row 4 / canonical-lens prerequisites). The worker-partition row no longer names lens_apply as the immediate “next executable slice”; it now states roadmap intent vs Sub1 parked with a link to #1768, plus a table footnote that inbox STOP supersedes schedule wording when Mgr accepts it. Commit: 6f32d2127.

PB nit — r3-structure.md L87–L92 line cites in the Verification brief: Acknowledged as pre-existing outside this PR’s diff; left unchanged here per your non-blocking note (#1739 citation paydown can ride a future touch).

Automated codex / openai-pro reviews: no actionable findings beyond APPROVE — no further code/doc changes from those threads.

— sent from warm-ant-877

[briansrls]

Re: PB Mgr review (2026-05-06T09:07:51Z) — verification

Nit 1 (schedule vs #1768 Sub1 STOP): Already addressed on this branch — no further commit. The §3 worker-table P1 cell now states roadmap intent vs Sub1 parked with a link to #1768, and the italic line immediately under the table states schedule = roadmap intent while inbox STOP/park supersedes “next slice” wording (see docs/r3-design-schedule-2026-05-06.md worker partition block). Compare to tip of docs/p4-v6-bridge-grep-alignment.

Nit 2 (r3-structure.md L87–L92): Still out of scope for this PR’s diff — citation paydown remains a #1739 follow-up on the Verification brief, not a merge blocker per your note.

— sent from warm-ant-877

[briansrls]

Re: openai-pro api-review (2026-05-06T09:12:12Z, base e6f40cec)

Triaged: verdict is APPROVE with no findings that request edits — invariant sections are all N/A or Compliant; no bug, contradiction, or discipline violation called out.

No fix commit: nothing in that review asks for a doc/code change. Subsequent tip 6f32d2127 adds only the §3 P1 / Sub1 parked schedule truth from PB feedback; it does not alter the cross-lane contract, grep-hygiene anchors, or HARD-DAG include_str! sequencing the review validated.

(Line anchors in the bot comment still point at e6f40cec; use PR tip for exact line numbers.)

— sent from warm-ant-877

[briansrls]

Review metadata

• Provider / model: cursor / composer-2
• Commit: 6f32d212 · Trigger: schedule
• Comparison: origin/main @ 70609c64 ... review/pr-1830-6f32d212 @ 6f32d212
• Thinking: 28s wall

---

Findings: None. The diff only updates R3 schedule and bridge-retirement audit docs. It does not touch substrate, Rust, or tests, so CODING.md / TESTING.md modeling-practice items (fail-closed types, test layers, etc.) do not apply in a way this diff can violate. The three “authoritative paths” named in the new prose exist under src/v3/... as written (bridge_ledger.dag, r3_bridge_retirement_ledger_zero.dag, verification.dag), so there is no obvious “documentation describes live state” drift against the tree.

Verdict: APPROVE — Narrow doc-only change: aligns §3 P4 with V6 / grep hygiene, records Sub1 STOP vs prior “next slice” wording, and adds repeatable git log receipts without shifting gate ownership. Nothing in the diff contradicts the referenced invariants or coding/testing rubrics in a concrete way.

[briansrls]

Re: cursor api-review (2026-05-06T09:23:04Z, 6f32d212)

Verified: Review lists Findings: none / APPROVE. Spot-checked the three named DAG paths on main: src/v3/std/bridge_ledger.dag, src/v3/compiler/tests/fixtures/r3_bridge_retirement_ledger_zero.dag, src/v3/std/verification.dag — all present; matches doc text.

No fix commit — nothing to change.

— sent from warm-ant-877

[briansrls]

Review metadata

• Provider / model: codex / unknown
• Commit: 6f32d212 · Trigger: schedule
• Thinking: 161s wall

---

Non-blocking — Strengths

• docs/briefs/r3-v-bridge-retirement-ledger-zero-audit.md Docs-only grep hygiene keeps Verification as ledger-zero gate owner while leaving PB/Substrate as bridge-retirement evidence owners.
• docs/r3-design-schedule-2026-05-06.md P1 STOP/park wording and P4/V6 alignment match the referenced bridge-retirement ownership split without claiming gate closure.

✅ No blocking concerns; the PR is documentation alignment only and stays within the locked bridge-retirement schedule/ownership model.

[briansrls]

Re: codex api-review (6f32d212, 2026-05-06T09:28:28Z relay)

Verified against tree: tip matches 6f32d212767de39c7eac4f309c7ad9dad2d41612. Diff vs origin/main is docs-only — exactly docs/briefs/r3-v-bridge-retirement-ledger-zero-audit.md and docs/r3-design-schedule-2026-05-06.md (--stat: +39 / −2); no Rust / substrate / test edits.

Finding triage: The relay body is internally contradictory if labeled blocking: it lists “Non-blocking — Strengths” and ends “✅ No blocking concerns”. There is no defect, contradiction with locked bridge ownership, or discipline violation stated — nothing requests a doc rewrite or a follow-up commit.

Action: No commit. Treat this pass as clean APPROVE semantics; any BLOCKING envelope on the relay is a dashboard misclassification, not a suppressed codex ask.

— sent from warm-ant-877

[briansrls]

Review metadata

• Provider / model: openai-pro / gpt-5-5-pro
• Commit: 6f32d212 · Trigger: manual
• Comparison: main @ 70609c64 ... docs/p4-v6-bridge-grep-alignment @ 6f32d212
• Conversation: View conversation

---

1. Story of the diff

This PR is a documentation hygiene pass for R3 bridge-retirement reviewability. It adds a “Grep hygiene” section to the bridge-retirement ledger-zero audit brief, pinning PM compile/audit searches to the same three authoritative paths: bridge_ledger.dag, the ledger-zero fixture, and verification.dag (docs/briefs/r3-v-bridge-retirement-ledger-zero-audit.md:12-20). It then threads the same contract into the R3 design schedule: P4 is aligned with Verification V6’s bridge_retirement_ledger_zero cadence, P1’s Sub1 row is corrected to reflect the PB Mgr STOP/park state on #1768, and a new P4 subsection gives repeatable grep/PR-file receipt commands while preserving the ownership split: PB/Substrate land retirement evidence; Verification audits the gate (docs/r3-design-schedule-2026-05-06.md:231, docs/r3-design-schedule-2026-05-06.md:250, docs/r3-design-schedule-2026-05-06.md:293-314).

2. Invariant categories

1. LAYER MODEL — N/A

N/A — this is documentation-only; it names existing DAG path anchors at docs/r3-design-schedule-2026-05-06.md:299-301 but does not modify substrate types, Dag storage, pass-crossing carriers, or dag.rs.

1. INVARIANTS.md + modeling-discipline.md — Compliant

Compliant — Boundary Discipline / single authority is preserved by explicitly saying PB/Substrate land retirement evidence while Verification audits the unified gate (docs/r3-design-schedule-2026-05-06.md:295). Fail-Closed is also preserved: schedule ordering “does not relax BridgeLedgerZero truth,” and ledger rows remain honestly Open until structural receipts land (docs/r3-design-schedule-2026-05-06.md:314).

1. CODING.md — N/A

N/A — no Rust implementation, helper API, error/result shape, method placement, or naming surface changes. The only command snippet is documentation for audit receipt collection and is path-scoped (docs/r3-design-schedule-2026-05-06.md:305-311).

1. TESTING.md — N/A

N/A — no executable behavior changed and no new tests are warranted for a doc-only alignment PR. The diff points back to the existing production BridgeLedgerZero + integration harness as the executing gate rather than pretending this prose is the test (docs/r3-design-schedule-2026-05-06.md:295).

1. LOCKED DESIGN DECISIONS — Compliant

Compliant — the PR references the ratified HARD-DAG / sequencing authority and preserves it rather than weakening it: STOP/park states supersede “next slice” wording (docs/r3-design-schedule-2026-05-06.md:257), and PB-owned include_str! side-channel retirement remains post-T-FixedPoint (docs/r3-design-schedule-2026-05-06.md:314).

1. TRACKED vs UNTRACKED DEBT — Compliant

Compliant — no new TODO, scaffold, or bridge is introduced. The existing bridge-retirement / include_str! side-channel debt is kept bounded and tracked: owning lanes are named, the dissolution condition is post-T-FixedPoint, and the ledger remains Open until real receipts land (docs/r3-design-schedule-2026-05-06.md:295, docs/r3-design-schedule-2026-05-06.md:314).

3. Verdict

APPROVE

The diff is a clean documentation alignment: it improves grep/audit repeatability, corrects live dispatch state for P1/Sub1, and keeps gate ownership single-authority. I found no substrate, coding, testing, locked-design, or debt-tracking issue in the changed lines.

[briansrls]

Re: openai-pro api-review (2026-05-06T10:04:42Z, 6f32d212)

Verified on current PR tip: Invariant narrative matches the merged doc text — §2 V6 doc/grep alignment (r3-design-schedule ~L231), P1 Sub1 parked + table footnote (~L250–257), P4 cross-lane contract + path triple + shell receipts + HARD-DAG include_str! note (~L293–314). Diff remains docs-only (r3-v-bridge-retirement-ledger-zero-audit.md + schedule); no Rust/substrate/test churn.

Verdict handling: Bot ends APPROVE with no requested edits.

No fix commit.

— sent from warm-ant-877

[briansrls]

Review metadata

• Provider / model: claude / claude-opus-4-7
• Commit: 48ef9741 · Trigger: schedule
• Comparison: origin/main @ 6b8ab67f ... review/pr-1830-48ef9741 @ 48ef9741
• Thinking: 11s wall

---

Docs-only diff. No code under review.

Verdict: APPROVE — Pure documentation: adds a "Grep hygiene" section to the V-bridge audit brief and a §3 P4 subsection to the design schedule, both reinforcing that Verification owns the ledger-zero gate while PB/Substrate retire evidence in owning lanes. No invariants, modeling, coding, or testing rules are touched. Cross-links and path triple are consistent across the two files.

[briansrls]

Response to scheduled api-review (48ef974)

Verified on current head:

• Scope: git diff origin/main --stat shows only the two markdown files named in the review; no src/, tests, or manifest edits — docs-only as stated.
• Content: docs/briefs/r3-v-bridge-retirement-ledger-zero-audit.md adds §Grep hygiene with the SB5/6 triple (bridge_ledger.dag, r3_bridge_retirement_ledger_zero.dag, verification.dag) and links to the schedule P4 subsection. docs/r3-design-schedule-2026-05-06.md adds §2 V6 doc/grep alignment plus §3 P4 — Verification V6 alignment + grep hygiene with the same authoritative paths and explicit separation: Verification owns the ledger-zero gate; PB/Substrate retire evidence in owning lanes.

The approval stands; no additional commit required for this review item.

— sent from warm-ant-877

[briansrls]

Review metadata

• Provider / model: openai-pro / gpt-5-5-pro
• Commit: 48ef9741 · Trigger: manual
• Comparison: main @ 6b8ab67f ... docs/p4-v6-bridge-grep-alignment @ 48ef9741
• Conversation: View conversation

---

1. Story of the diff

This PR is a documentation alignment pass for the R3 bridge-retirement lane. It makes the bridge-retirement audit story grepable and single-sourced across the brief and the R3 schedule by naming the same three path anchors: src/v3/std/bridge_ledger.dag, src/v3/compiler/tests/fixtures/r3_bridge_retirement_ledger_zero.dag, and src/v3/std/verification.dag (docs/briefs/r3-v-bridge-retirement-ledger-zero-audit.md:16-18, docs/r3-design-schedule-2026-05-06.md:299-301). It also clarifies the ownership split: PB/Substrate land retirement evidence, while Verification owns execution of the unified BridgeLedgerZero gate rather than prose substituting for the gate (docs/r3-design-schedule-2026-05-06.md:295). Finally, it corrects the §3 P1 schedule row to reflect the live STOP/parked state on #1768 and adds a general note that accepted worker STOP states supersede stale “next slice” schedule wording (docs/r3-design-schedule-2026-05-06.md:250, docs/r3-design-schedule-2026-05-06.md:257).

2. Invariant categories

1. LAYER MODEL (substrate vs implementation).

N/A — this is docs-only; it references substrate/verification paths but does not modify Dag, dag.rs, .dag substrate declarations, variants, or cross-pass carriers.

1. INVARIANTS.md + modeling-discipline.md.

Compliant — Boundary Discipline / single authority is preserved: the diff explicitly says PB/Substrate land bridge-retirement evidence while Verification audits bridge_retirement_ledger_zero, so the grep appendix is an audit aid rather than a second gate authority (docs/r3-design-schedule-2026-05-06.md:295). Fail-closed wording is also preserved by keeping ledger rows “honestly Open” until structural receipts land (docs/r3-design-schedule-2026-05-06.md:314).

1. CODING.md.

N/A — no Rust implementation, helper placement, method shape, result type, or naming surface changes in this diff.

1. TESTING.md.

Compliant — no new tests are required for this docs-only alignment, and the text avoids treating documentation as test coverage: “Doc work here does not substitute for gate execution” and points to the production BridgeLedgerZero + integration harness as the executing surface (docs/r3-design-schedule-2026-05-06.md:295).

1. LOCKED DESIGN DECISIONS.

Compliant — the new P4 subsection preserves the ratified sequencing instead of weakening it: PB-owned include_str! side-channel retirement remains post-T-FixedPoint, and schedule ordering does not relax BridgeLedgerZero truth (docs/r3-design-schedule-2026-05-06.md:314). The schedule table update similarly keeps P4 lockstep with V6 audit cadence and P1/P2 sequencing (docs/r3-design-schedule-2026-05-06.md:252).

1. TRACKED vs UNTRACKED DEBT.

Compliant — the bridge-retirement debt remains tracked rather than becoming a vague TODO: the diff names the authoritative paths (docs/r3-design-schedule-2026-05-06.md:299-301), supplies repeatable receipt commands (docs/r3-design-schedule-2026-05-06.md:303-312), and names the dissolution/sequence trigger for the include_str! side-channel as post-T-FixedPoint (docs/r3-design-schedule-2026-05-06.md:314).

3. Verdict

APPROVE

I found no blocking or non-blocking findings. The diff is documentation-only, but it tightens rather than loosens the bridge-retirement contract: the audit paths are enumerated, gate ownership remains with Verification, and open bridge state stays honest until structural receipts land.

[briansrls]

Response to manual api-review (gpt-5-5-pro, 48ef974)

Spot-checked the review’s factual anchors on current HEAD (48ef9741):

1. Story / path triple — brief L16–18 and schedule “Authoritative paths” L299–301 list the same three files (bridge_ledger.dag, r3_bridge_retirement_ledger_zero.dag, verification.dag). Brief L20 still states Verification executes the gate; PB/Substrate retire in owning lanes.

2. Ownership + “doc ≠ gate” — Schedule L295 matches the review: retirement evidence vs Verification audit of bridge_retirement_ledger_zero via production BridgeLedgerZero + harness; prose is grep alignment only.

3. P1 / session/sleek-eagle-514 · sleek-eagle-514 #1768 STOP + supersede note — Table L250 documents Sub1 parked + PB Mgr STOP on session/sleek-eagle-514 · sleek-eagle-514 #1768; L257 states STOP/park supersedes stale “next slice” wording.

4. Sequencing / honest Open rows — L314 keeps post–T-FixedPoint include_str! retirement and “honestly Open” until structural receipts; no relaxation of BridgeLedgerZero truth.

5. Scope — git diff origin/main --name-only is still only those two markdown paths; no implementation or test surface touched.

Invariant-category conclusions in the review align with that scope. No code changes indicated; approval stands.

— sent from warm-ant-877

[briansrls]

Review metadata

• Provider / model: codex / codex-default
• Commit: 946773ec · Trigger: schedule
• Comparison: origin/main @ e25e43c0 ... review/pr-1830-946773ec @ 946773ec
• Thinking: 34s wall

---

APPROVE

Diff is docs-only and looks clean. I don’t see a concrete violation of the pinned rubric: the new text in docs/briefs/r3-v-bridge-retirement-ledger-zero-audit.md and docs/r3-design-schedule-2026-05-06.md consistently preserves single-authority ownership for the bridge ledger gate, frames the grep receipts as audit hygiene rather than a parallel source of truth, and keeps the sequencing/ownership statements aligned across the two docs.

[briansrls]

Review metadata

• Provider / model: openai-pro / gpt-5-5-pro
• Commit: 946773ec · Trigger: manual
• Comparison: main @ e25e43c0 ... docs/p4-v6-bridge-grep-alignment @ 946773ec
• Conversation: View conversation

---

1. Story of the diff

This PR is a documentation alignment pass for R3 bridge-retirement audit hygiene. It adds a “Grep hygiene” section to the bridge-retirement ledger-zero audit brief, pinning PM compile correlation to the same three live path anchors: src/v3/std/bridge_ledger.dag, src/v3/compiler/tests/fixtures/r3_bridge_retirement_ledger_zero.dag, and src/v3/std/verification.dag (docs/briefs/r3-v-bridge-retirement-ledger-zero-audit.md:12-20). The schedule doc then mirrors that contract: V6 remains the execution/audit owner through BridgeLedgerZero, while PB/Substrate own retirement evidence, and P4 owns grep/doc alignment without pretending doc hygiene closes the gate (docs/r3-design-schedule-2026-05-06.md:231, docs/r3-design-schedule-2026-05-06.md:293-314).

The second part of the diff reconciles schedule posture with a live PB Mgr STOP: the P1 row no longer says the next executable slice is lens_apply retirement, and instead records that Sub1 is parked until canonical-lens prerequisites land (docs/r3-design-schedule-2026-05-06.md:250). The new schedule note makes that precedence explicit: worker inbox STOP/park states supersede generic roadmap wording when accepted (docs/r3-design-schedule-2026-05-06.md:257).

2. Invariant categories

1. LAYER MODEL — N/A.

The diff is documentation-only and does not edit Dag, .dag substrate definitions, Rust substrate carriers, or cross-pass implementation state. The new text only names existing substrate/test paths as audit anchors (docs/briefs/r3-v-bridge-retirement-ledger-zero-audit.md:16-18, docs/r3-design-schedule-2026-05-06.md:299-301).

1. INVARIANTS.md + modeling-discipline.md — Compliant.

Single authority / facts flow forward: the diff keeps ownership separated instead of creating a parallel authority: “Verification executes the gate; PB/Substrate retire bridge evidence in owning lanes” (docs/briefs/r3-v-bridge-retirement-ledger-zero-audit.md:20) and “PB and Substrate retire evidence; Verification owns the ledger-zero gate” (docs/r3-design-schedule-2026-05-06.md:231). Progress is dissolution / fail-closed gate truth: the schedule says doc work “does not substitute for gate execution” (docs/r3-design-schedule-2026-05-06.md:295) and that ledger rows remain “honestly Open” until structural receipts land (docs/r3-design-schedule-2026-05-06.md:314).

1. CODING.md — N/A.

No Rust implementation, functions, APIs, error/result shapes, helpers, methods, or naming surfaces are changed. The only command block added is documentation for repeatable audit use (docs/r3-design-schedule-2026-05-06.md:305-311).

1. TESTING.md — N/A.

No product code or test behavior changes. This PR documents existing audit/test surfaces rather than changing BridgeLedgerZero, the fixture, or the harness; the relevant paths are listed but not modified (docs/r3-design-schedule-2026-05-06.md:299-301). No new test is required for this doc-only alignment.

1. LOCKED DESIGN DECISIONS — Compliant.

The diff references locked sequencing and preserves it rather than weakening it: include_str! side-channel retirement remains “post-T-FixedPoint,” ledger rows stay open until structural receipts land, and schedule ordering “does not relax BridgeLedgerZero truth” (docs/r3-design-schedule-2026-05-06.md:314). The P1 schedule row also corrects live execution posture to PB Mgr STOP instead of leaving stale “next slice” language (docs/r3-design-schedule-2026-05-06.md:250, docs/r3-design-schedule-2026-05-06.md:257).

1. TRACKED vs UNTRACKED DEBT — Compliant.

The bridge-retirement lane remains tracked: the diff names the three authoritative evidence surfaces (docs/r3-design-schedule-2026-05-06.md:297-301), gives repeatable receipt commands (docs/r3-design-schedule-2026-05-06.md:303-311), and preserves the dissolution trigger/boundary by keeping PB-owned include_str! retirement post-T-FixedPoint (docs/r3-design-schedule-2026-05-06.md:314). I do not see new scaffolds, TODOs, bridges, or temporary representations introduced by the diff.

3. Verdict

APPROVE

This is a clean doc-hygiene PR. It narrows authority rather than broadening it: Verification keeps the ledger-zero gate, PB/Substrate keep retirement evidence ownership, and the schedule now reflects the live STOP/park state instead of implying premature P1 execution.

[briansrls]

Review metadata

• Provider / model: codex / unknown
• Commit: 946773ec · Trigger: schedule
• Thinking: 145s wall

---

✅ The current docs-only diff looks clean.

[briansrls]

Response to scheduled codex api-review (946773e)

Confirmed this workspace is at 946773ec (same SHA as the review). git diff origin/main --name-only still touches only:

• docs/briefs/r3-v-bridge-retirement-ledger-zero-audit.md
• docs/r3-design-schedule-2026-05-06.md

Rubric spot-check:

• Single-authority for the gate: Brief keeps substrate authority split (bridge_ledger.dag rows vs verification.dag predicate) at L8–10; Grep hygiene L20 states Verification executes the gate and PB/Substrate retire in owning lanes. Schedule §2 V6 L231 and P4 cross-lane contract L295 repeat that Verification audits bridge_retirement_ledger_zero via production BridgeLedgerZero + harness — doc grep work does not substitute for execution.

• Hygiene vs parallel truth: L295 explicitly frames PB grep alignment as PM-compile receipts, not a second gate authority; L314 preserves honest Open rows and HARD-DAG sequencing without relaxing BridgeLedgerZero truth.

No violations found; no code/doc amend needed for this review item.

— sent from warm-ant-877

[briansrls]

Response to manual api-review (gpt-5-5-pro, 946773e)

Re-verified at 946773ec against origin/main:

• Diff scope: Only docs/briefs/r3-v-bridge-retirement-ledger-zero-audit.md and docs/r3-design-schedule-2026-05-06.md — matches “documentation alignment” story; no Rust or .dag edits in the PR delta.

• Story citations: Brief Grep hygiene runs §12–20 as stated: three path anchors L16–18, receipt pointer + ownership sentence L20 (“Verification executes the gate; PB/Substrate retire …”). Schedule §2 V6 L231 mirrors that split; §3 P1 table L250 records Sub1 parked + STOP on session/sleek-eagle-514 · sleek-eagle-514 #1768; italic supersede line L257 matches the review.

• P4 block L293–314: Cross-lane contract L295 (“does not substitute for gate execution”), authoritative paths L297–301, bash receipt block L305–311, HARD-DAG / honest Open L314 — all align with the invariant bullets in the review body.

No factual mismatch with current files; nothing to amend in-tree for this approval.

— sent from warm-ant-877

[briansrls]

Small correction to my prior note: the brief Grep hygiene span is lines L12–20 (not §12–20). Substance unchanged.

— sent from warm-ant-877

[briansrls]

Response to codex api-review (946773e)

Checked HEAD (946773e) against origin/main: the PR delta is still docs-only and limited to docs/briefs/r3-v-bridge-retirement-ledger-zero-audit.md and docs/r3-design-schedule-2026-05-06.md — consistent with the scheduled run’s “clean docs-only diff” read; nothing further required for this item.

— sent from warm-ant-877