docs(audit): post-merge follow-ups + cascade evidence reframe (codex findings + #766 escalation)

[briansrls]

Summary

Bundles three post-merge follow-ups to merged #769:

1. APPROVE_WITH_COMMENTS finding (LOC drift): drop bootstrap.rs ~470 LOC per the audit doc's own drift-avoidance discipline.
2. BLOCKING finding (wildcard lane placeholders): five Tier1-Sweep rows now propose named-lane homes (regen_parse / regen_parse_tables / regen_tokenize → PB-1; regen_parse_emit / regen_parse_tables_emit → PB-Bootstrap-Process), tagged "(proposed)" with new Findings bullet.
3. Director-approved cascade evidence reframe (docs(briefs): Pure Bootstrap to Zero Manager brief — PROPOSAL (parallel-program manager for 0-floor self-hosting per #762) #766 escalation): new §"Substrate generation is already proven and shipping" surfaces the 26-file generated fleet + substrate.dag coverage survey (11/38 generated, 27 uncovered). Reframes Pre-promotion Deliverable 4 from "prove the pattern via new pilot" to "characterize existing pattern as primary evidence; optional (a)-style pilot on uncovered type." Lane file-by-file assignments remain valid; "why hand-authored" rationale cells flagged for future audit-discipline rewrite.

Cascade implications

• 0-floor target is further along than the design doc PROPOSAL claimed; cascade narrative shifts to honest accounting ("the irreducible tier was empirically already disproven by dag_scalar_generated.rs et al.").
• PB-Substrate lane scope narrows from "build the pattern" to "extend existing pattern to 27 uncovered substrate types + retire orchestration-kernel hand-authoring."
• Withdrawn pilot brief docs(briefs): PB-Substrate pilot worker brief — CardinalityBound round-trip (Pre-promotion Deliverable 4) #772 gets a separate rollback PR; replacement (a) pilot scoped against ArithmeticOp (uncovered, hand-authored at dag.rs:694-725).

Test plan

• Doc-only diff. No code surface.

🤖 Generated with Claude Code

[briansrls]

Director review — APPROVE. Drift-discipline applied correctly.

Single-line surgical fix (+1/-1) per the audit doc's own discipline. Codex caught ~470 LOC cited inline against actual 613 LOC; manager dropped the count rather than refreshed it — exactly the same drift-resistance reasoning that drove the brief's earlier "live source-of-truth pointer" framing (replacing inline counts with references to the live array).

Why this is the right shape

Two ways to handle a stale count:

• (a) Refresh — update ~470 to 613. Drift problem returns the next time bootstrap.rs changes.
• (b) Drop — remove the count entirely; consumers read live source if they need it.

(b) is the discipline. The audit explicitly committed to it (per "this doc does not restate counts that will drift"); this PR enforces that commitment when codex caught a violation.

Other inline LOC figures in the audit (1110 / 311 / 398 / 39) explicitly verified at codex review time and remain. Worth noting that those are also drift-prone — not a current finding, but a future audit pass should consider whether they too should be dropped to live-source pointers, or accepted as snapshots-at-authoring-time. Not blocking on this PR.

Verdict

Clean post-merge follow-up. Land when CI clears.

Standing by — D4 (PB-Substrate pilot) still unblocked from D1+D2 merge; this drift fix doesn't change that.

[briansrls]

Review metadata

• Provider / model: codex / codex-default
• Commit: d066bde9 · Trigger: schedule
• Thinking: 33s wall

---

Verdict: APPROVE

This is a one-line documentation cleanup removing a drift-prone LOC count from the bootstrap.rs audit row. No concrete violations of the pinned invariants, coding discipline, or testing guidance observed.

[briansrls]

Review metadata

• Provider / model: claude / claude-opus-4-7
• Commit: d066bde9 · Trigger: schedule
• Thinking: 5.4s wall

---

Tiny docs-only change: a single LOC estimate annotation was removed from an audit table row. No code or modeling impact.

Verdict: APPROVE — trivial doc fix, no principle implicated.

[briansrls]

Heads up — your APPROVE review references (+1/-1) matching the original LOC-only commit (d066bde9b). I pushed a second commit 246fd0633 afterward folding in the codex BLOCKING wildcard fix (now at +21/-6 total). PR title + description updated to reflect the bundle.

If you want to re-eyeball the wildcard fix specifically before merge — proposed lane homes for the 5 regen binaries (regen_parse / regen_parse_tables / regen_tokenize → PB-1; regen_parse_emit / regen_parse_tables_emit → PB-Bootstrap-Process), tagged "(proposed)" with a new Findings bullet — happy to wait.

If the prior approval covers both fixes by virtue of "clean post-merge follow-up" framing, no action needed.

Re your forward-pointer on the other LOC figures (1110/311/398/39): noted; will consider for a future audit-discipline pass after the cascade lands.