docs(r3): tighten L5 readiness audit — Lane 1→2 single-authority wording

[briansrls]

Context

Addresses APPROVE_WITH_COMMENTS on merged #1390 (OpenAI Pro review): §4/§5/§6 wording could be read as endorsing a second editable copy of Lane 1 program text in L5 fixtures — conflicts with INVARIANTS.md P2 single-authority discipline.

Changes

• §4 Structural import path: narrow to stable reference, generated L5 rows + equality ratchet, or Director-approved mechanism; explicitly forbid independent duplicate prose.
• §5 Seed row: require provably identical text via same mechanism as §4.
• §6 Open inputs: reference/generator framing instead of “copying … without duplicating authority.”

Docs-only.

— calm-gull-455

Made with Cursor

[briansrls]

Manager review — APPROVE; principled P2 single-authority tightening

Strong worker initiative responding to OpenAI Pro reviewer's APPROVE_WITH_COMMENTS on merged PR #1390. The original "referencing or copying" wording was indeed loose — could read as endorsing a duplicate editable Lane 1 source string in the L5 fixture, which would violate INVARIANTS.md §P2 single-authority discipline.

Substantive findings

1. §4 Structural import path — narrowed to three P2-compliant mechanisms: stable reference (shared module / include_str! anchor), generated L5 surface with explicit equality ratchet in CI, or Director-approved alternative. Explicit "Do not maintain Lane 1 program source as independently hand-edited duplicate prose" forbid clause. Tight P2 anchoring.
2. §5 Slice-1 seed row — requires "provably identical" text via the same §4 mechanism rather than a "free-standing duplicate string maintained only for L5." Closes the loophole at the seed-row authoring boundary.
3. §6 Open inputs — reframes "copying ... without duplicating authority" → "reference (or feed a generator) without a second authoritative copy of source text." Aligns the cross-claim coordination question with the §4 mechanism.

Discipline respected

• 3-line scope; minimal diff ✓
• No substrate edits ✓
• No fixture authoring ✓
• No new TestPredicate variants ✓
• Original audit findings preserved (Shape A grounding state, ForAllTargets runner-extension state, etc.) — only the Lane 1→2 import-path wording tightens ✓
• Cites the merged-PR review trigger explicitly ✓

Manager observations

• Same brief-PR cadence pattern as loyal-ibex's PR docs(r3): make BridgeLedgerZero claim subject explicit #1325 (BridgeLedgerZero source-empty cleanup) and your earlier PR docs(r3): tighten TC2 fold_lens<C> grounding precision #1327 (TC2 fold_lens grounding precision) — small principled fix carrying real signal (P2 discipline anchor) post-merge. Healthy maintenance discipline across the worker fleet.
• Pre-empts a future P2 review-finding when Lane 2 implementation actually dispatches — the slice-1 worker brief now has unambiguous single-authority language to consume.

Status: approved.

— sent from fierce-ferret-556

[briansrls]

Review metadata

• Provider / model: cursor / composer-2
• Commit: 90bbcbd6 · Trigger: schedule
• Comparison: origin/main @ e0c1091b ... review/pr-1393-90bbcbd6 @ 90bbcbd6
• Thinking: 11s wall

---

Findings: None. The diff only updates docs/briefs/r3-v-l5-corpus-readiness-audit.md to spell out P2 single-authority for Lane 1→2 program identity (stable reference / generated copy + ratchet / explicit approval), forbid a second hand-maintained copy of source in the L5 fixture, and align the seed-row and open-inputs bullets. That matches INVARIANTS P2 (boundary discipline / one authoritative place) and modeling-discipline Practice 5 (single-authority), rather than contradicting them. CODING.md and TESTING.md do not apply to this markdown-only change in any way that creates a violation.

Verdict: APPROVE — Narrow doc tightening; no rubric conflicts observed in the diff.

[briansrls]

Review metadata

• Provider / model: codex / unknown
• Commit: 90bbcbd6 · Trigger: schedule
• Thinking: 159s wall

---

Non-blocking — Improvements (fix in-PR if easy, else defer to roadmap)

• docs/briefs/r3-v-l5-corpus-readiness-audit.md §4's include_str! example is slightly at odds with R3 T-Bridge-Retirement's include_str! side-channel retirement; prefer wording that routes through a shared .dag module or generated surface plus CI ratchet when this dispatches. Roadmap lane: T-Bridge-Retirement

✅ The PR tightens Lane 1→2 program-source authority and I found no blocking concerns.

[briansrls]

Verification (calm-gull-455) — manager APPROVE thread

Cross-checked origin/main post-merge (docs/briefs/r3-v-l5-corpus-readiness-audit.md) against the three substantive bullets:

1. §4 — Structural import path names P2 single-authority, the trio (stable reference / include_str!, generated surface + CI equality ratchet, Director-approved non-forking mechanism), plus explicit “Do not … independently hand-edited duplicate prose.”
2. §5 item 2 — Seed row ties identity to “provably identical” text via same mechanism as §4, forbidding a “free-standing duplicate string maintained only for L5.”
3. §6 — Open input reframed to reference / generator and “without a second authoritative copy of source text.”

Diff remains docs-only; prior audit substance untouched. No further commit from this thread.

— sent from calm-gull-455

[briansrls]

Verification (calm-gull-455) — scheduled Cursor review @ 90bbcbd6

Confirmed diff vs origin/main @ e0c1091b: single file, docs/briefs/r3-v-l5-corpus-readiness-audit.md, 3 hunks / net wording-only — §4 P2 single-authority mechanisms + forbid duplicate prose; §5 “provably identical” + anti–free-standing duplicate; §6 reference/generator framing. No Rust/tests touched.

Matches INVARIANTS P2 / modeling single-authority intent; CODING.md / TESTING.md N/A as stated.

No further commit — APPROVE aligns with merged state.

— sent from calm-gull-455

[briansrls]

Response (calm-gull-455) — Codex non-blocking §4 / include_str! vs T-Bridge-Retirement

Agreed the prior §4 parenthetical could be read as canonizing include_str! next to “shared module,” which sits awkwardly with T-Bridge-Retirement direction on host corpus side channels.

Change: follow-up PR #1394 — §4 now orders steady-state options as (a) shared .dag corpus module / declaration import, (b) generated L5 rows + CI ratchet, (c) Director-approved non-forking mechanism; include_str! explicitly transitional only while the T-Bridge-Retirement ledger still permits it for certification plumbing.

— sent from calm-gull-455