lane-asks/INDEX.md

# Lane Ask Queue — ZPE-Neuro

## Identity
- Lane: `ZPE-Neuro`
- Repo: `github.com/Zer0pa/ZPE-Neuro` (PUBLIC)
- Section: `encoding`
- Website .md: `src/content/lanes/encoding/ZPE-Neuro.md`
- Website URL: http://127.0.0.1:4500/encoding/ZPE-Neuro/
- Brief (canonical): `/Users/Zer0pa/orchestration-state/LANE_AGENT_BRIEF_2026-05-03.md`

## Known flags
(none)

---

## Ask log (newest first; append above the divider)

### 2026-05-08 · Wave-2.1 headline outsider-readability fix (Corridor Agent)

**Trigger:** Wave-2 re-dispatch (2026-05-08 ~14:21 SAST) added step 2a — "the headline must make sense on first read to a smart outsider: it must say what the lane does, what it is trying to do, or what failed." Previous headline `Spikes you can replay bit-exact.` failed this test: "Spikes" alone is ambiguous (electrical, statistical, neural) and the cat-phrase `· ZPE-NEURO · SPIKE EXTRACTION` was carrying the domain context — exactly the failure mode the brief warns against.

**Fix:** prepend `Brain` so the domain anchor lives in the headline itself. Accent span unchanged (load-bearing claim still on `replay bit-exact.`).

- **Before:** `<h2 class="lane-headline">Spikes you can <span class="accent">replay bit-exact.</span></h2>` — 5 words, 32 chars.
- **After:**  `<h2 class="lane-headline">Brain spikes you can <span class="accent">replay bit-exact.</span></h2>` — 6 words, 38 chars. Within §C.1 4–7 word and §C.2 25–40 char envelopes.

**Smart-outsider test (step 2a):** "BRAIN SPIKES YOU CAN REPLAY BIT-EXACT." — domain (brain) + object (spikes) + claim (replay bit-exact). Outsider parses: this is about neural spike events that can be reproduced exactly. The domain anchor no longer depends on the cat-phrase or body to disambiguate. Headline does not end in any of the brief's flagged endings (`closed chain`, `owned pending`, `intake dossier`, `structural audit`, `authority packet`, `declared ground`, `each exact`); "bit-exact" is a real software/audio engineering term, not internal jargon.

**Truth-envelope check on the fix:** §A non-claim boundary says "must NOT imply broad neural-decoding coverage, on-silicon measurement, or commercialization-ready surface. Bound to DANDI 000034 spike-event extraction." The new headline names the data domain (brain spikes) but makes no decoding/coverage/silicon/commercial claim — the claim is determinism (replay bit-exact). Body still bounds the surface to DANDI 000034 verbatim. Blocker still names DANDI 000003 breadth FAIL + modeled-not-on-silicon + no-broad-coverage. §A boundary holds.

**Block touched:** the `<h2 class="lane-headline">…</h2>` element inside `<article data-lane="zpe-neuro">` only. One atomic Python rewrite (1 char insertion of `Brain ` before `spikes`). No `data-status` change, no scaffold change, no SVG change, no metric change, no blocker change, no CTA change, no neighbouring-block change.

**Files changed (this session):** `Website-fresh/.claude/worktrees/landing-rd-v3/site-v2/src/landing-rd-v3.template.html` — single `<h2>` text edit. No other files in any worktree, repo, scaffold, CSS, or JS were touched.

**Receipts captured at `/Users/Zer0pa/orchestration-state/receipts/ZPE-Neuro-2026-05-08-wave2.1/`:**
1. `desktop-1440-block.png` — block-only at 1440 / dsf=1, count-up animation completed (`401×`, `78.44 µV` resolved, not the in-flight `374` / `73.68` mid-animation values). Headline wraps to 2 lines: `BRAIN SPIKES YOU CAN` / `REPLAY BIT-EXACT.` Body, metrics, viz with caption (no overlap), blocker (trimmed 235-char version), and `OPEN REPO SURFACE →` CTA all rendering cleanly.
2. `desktop-1440-context.png` — slot 17 row with peer slot 16 ZPE-IoT (`SENSOR TELEMETRY, TEN OF TEN WINS.` — also 2-line headline; rhythm matches), plus tail of row above and start of row below (slot 18 ZPE-Taste + slot 19 ZPE-Geo). Card height matches row peer; grid rhythm holds.
3. `mobile-414-block.png` and `mobile-414-context.png` — at 375–414 mobile fold, `.lanes-grid` collapses to single column at `max-width: 980px` and `.lane-body` collapses at `max-width: 720px`. Headline wraps to 3 lines on mobile (`BRAIN SPIKES YOU` / `CAN REPLAY BIT-` / `EXACT.`); body paragraph stacks above metrics; metrics stack above viz. Card reads cleanly at mobile fold; no truncation; no overflow. Mobile receipts now landed (the Wave-2 packet noted these were blocked by intermittent dev-server 500 + Chrome tile-memory limits; this re-dispatch retook them via Playwright with longer count-up wait).

**Benchmark-footprint and style compliance (re-verified post-headline-fix):**
- Headline: 6 words / 38 chars, accent on `replay bit-exact.` ending in period, indicative present tense, no hedge. Within §C.1 envelope. Wraps to 2 lines on desktop / 3 lines on mobile (consistent with row peer ZPE-IoT and reference ZeroShip wrap behaviour).
- Lane-tag cat: unchanged — `· ZPE-NEURO · SPIKE EXTRACTION` (4 words, 30 chars).
- Human paragraph: unchanged — 342 chars, 58 words, one `<strong>` span.
- Metrics: unchanged — 4 rows (`401×` / `78.44 µV` / `5 / 5 SEEDS` / `C & D PASS`).
- Viz: unchanged — 3-section schematic, ~3.0 KB markup, viewBox 0 0 400 260, white-only strokes, three callouts.
- Caption: unchanged — `FIG · DANDI 000034 · BIT-EXACT SPIKE EXTRACTION` (47 chars, within §C.3 ≤50; wraps to 2 lines at the rendered viz column width — within envelope, consistent with several reference cards).
- Blocker: unchanged from Wave-2 trim — 235 chars, 35 words.
- CTA: unchanged — `OPEN REPO SURFACE` → `https://github.com/Zer0pa/ZPE-Neuro`.
- Scaffold: 4 corner spans on the article, 4 corner spans on the viz, `class="svg-soft"`, `data-repo="zpe-neuro"`, `viz-cap` div all present. No drift.

**Repo / web truth-sync performed this Wave-2.1 pass:** none. Wave-2 §B for ZPE-Neuro = `ALIGNED — No repo-front-door correction required from this pass.` Repo HEAD remains `3e6d53b`. Earlier-surfaced repo-side asks (G4 GitHub description, G4 topics overrun, G1 visibility/release row, README license-badge mismatch, "Honest Blocker INC" prose drift, stale `commit_sha`) remain surfaced-but-not-landed; they are operator scope under a separate repo-front-door pass.

**Resistance risks checked and not triggered:**
1. **Resistance 1 — Benchmark-shape mimicry:** card carries lane-specific cargo (DANDI 000034 / 401× / 78.44 µV / 5/5 seeds / Gate C/D / DANDI 000003 FAIL / IBL single-target / modeled-not-on-silicon). Removing the data leaves no recognisable benchmark-shape shell. Not triggered.
2. **Resistance 2 — Rush to green-flag:** the previous Wave-2 packet declared the lane fixed; the re-dispatch added a new criterion (step 2a smart-outsider test). Honoured the new criterion rather than declaring done. Not triggered.
3. **Resistance 3 — Approval via packet quality:** the page is the review surface; receipts cite the rendered card with count-up resolved. No polished-summary substitution for the actual fix. Not triggered.
4. **Resistance 4 — Efficiency as ownership fracture:** §A truth envelope, §B repo state, and §C UX layer all re-checked for this lane; continuity intact end-to-end after the fix. Not triggered.
5. **Resistance 5 — Flattery as freedom:** no bold rewrite. Single 1-word prepend (`Brain `), accent and load-bearing structure preserved. Not triggered.

`Lane ZPE-Neuro: wave-2 outcome = fixed (single bounded headline-domain-anchor fix per step 2a smart-outsider test); repo/front-door status = unchanged (§B = ALIGNED); landing-card status = populated; receipts ready.`

---

### 2026-05-08 · Landing-card Wave-2 polish complete (Corridor Agent)

**Wave:** Landing-card Wave-2 — inspect / falsify / fix / polish pass per `/Users/Zer0pa/orchestration-state/LANDING_CARD_WAVE2_CENTRAL_DISPATCH_BRIEF_2026-05-08.md`.

**Wave-2 class for this lane:** `WAVE2-CLEAN` per §A row "ZPE-Neuro — slot 17" (no truth drift, ownership intact, lane queue marks 2026-05-08 slot-17 POPULATED with receipts) + §B "ZPE-Neuro · ALIGNED · No repo-front-door correction required from this pass." Single bounded geometry-fix surfaced by §C.2 / §C.3 / §C.4: blocker 293 chars (over 280 ceiling by 5%); §C.4 read = `Hits / Hits / Approaches`.

**Block touched:** `<article data-lane="zpe-neuro" data-portfolio="encoding" data-status="populated">` in `/Users/Zer0pa/Website-fresh/.claude/worktrees/landing-rd-v3/site-v2/src/landing-rd-v3.template.html`. The article block itself, the preceding `<!-- CARD: -->` marker, the `data-lane` / `data-portfolio` / `data-status` attributes, the `<span class="num">17</span>`, the `<span class="stamp-block portfolio">ENCODING</span>` chip, and the `<span class="cat">…</span>` were all left untouched. The single Wave-2 edit is to the `.lane-blocker` prose text only.

**Files changed (this session):** `Website-fresh/.claude/worktrees/landing-rd-v3/site-v2/src/landing-rd-v3.template.html` — single `.lane-blocker` text edit via atomic Python rewrite (concurrent-edit safe). No other files in any worktree, repo, scaffold, CSS, or JS were touched.

**The fix — blocker compression (293 → 235 chars, 35 words):**
- **Before:** `DANDI 000003 breadth probe recorded FAIL; only one IBL-class extracellular target counted as breadth. Latency is a modeled cycle figure, not measured on silicon. No claim of broad neural-decoding coverage or commercial readiness beyond the bounded DANDI 000034 surface.`
- **After:** `DANDI 000003 breadth probe FAIL; only one IBL-class target counted. Latency is a modeled cycle figure, not measured on silicon. No claim of broad neural-decoding coverage or commercial readiness beyond the bounded DANDI 000034 surface.`
- **Disclaimers preserved verbatim:** (1) DANDI 000003 breadth probe FAIL; (2) IBL-class single-target breadth attribution; (3) latency is modeled-cycle, not measured on-silicon; (4) no broad neural-decoding coverage; (5) no commercial readiness; (6) bounded DANDI 000034 surface. All §A non-claim load-bearing facts kept.
- **Trim mechanism:** dropped "recorded" filler verb (-9 chars) and the "extracellular ... as breadth" tail clause (-26 chars); the IBL-class anchor is sufficient given the surrounding context (DANDI breadth probe). Total saved: 58 chars. New length 235 chars sits inside the §C.1 25–40-word / §C.2 ≤280-char envelope with margin.

**Truth basis used:**
- §A Slot 17 controlling proofs (`reports/ZPE-Neuro.md:24-34,51`; `falsification/ZPE-Neuro.md:35-92`; release v0.1.1; DANDI `000034`) — unchanged from Wave-1; §A reinspection re-confirmed verbatim.
- Wave-2 §A row for slot 17: "No drift. Latency disclosed as 'modeled cycle figure, not measured on silicon' per §A line 161. Breadth FAIL (DANDI 000003) named in blocker." — fix preserves both load-bearing facts.
- Wave-2 §B row: `ALIGNED · No repo-front-door correction required from this pass · Repo/card continuity is intact.` — no repo edits attempted; G1/G4 from earlier session remain surfaced-but-not-landed and are now explicitly out of Wave-2 scope per §B.
- Wave-2 §C.3 Slot 17 brief: viz archetype = substrate (already implemented), headline-accent direction = bounded-extraction (already implemented as "replay bit-exact" anchored to DANDI 000034 in body), body-copy pattern = authority-packet (already implemented), geometry note = "Blocker 293 chars. Trim one sentence to land in ≤280." — fix executes the geometry note.

**Resistance risks checked and not triggered:**
1. **Resistance 1 — Benchmark-shape mimicry:** card carries lane-specific cargo (DANDI 000034 / 401× / 78.44 µV / 5/5 seeds / Gate C/D / DANDI 000003 FAIL / IBL single-target / modeled-not-on-silicon). Removing the data leaves no recognizable benchmark-shape shell. Not triggered.
2. **Resistance 2 — Rush to green-flag:** §A flagged WAVE2-CLEAN but §C still surfaced a geometry signal; honoured the §C signal rather than declaring done. Trim is content-bound (§A facts preserved), not cosmetic. Not triggered.
3. **Resistance 3 — Approval via packet quality:** the page is the review surface; receipts here cite the in-source state and the rendered crop. No polished-summary substitution for the actual fix. Not triggered.
4. **Resistance 4 — Efficiency as ownership fracture:** Wave-1 + Wave-2 §A + §B + §C all checked end-to-end for this lane; continuity from repo-front-door (HEAD `3e6d53b`) → §A truth envelope → card prose is intact. Not triggered.
5. **Resistance 5 — Flattery as freedom:** no bold rewrite of headline / paragraph / metrics / viz / CTA attempted. Bounded fix to the one geometry signal §C named. Not triggered.

**Diff scope assertion:** the only Wave-2 edit is the `.lane-blocker` prose substring inside the `<article data-lane="zpe-neuro">` block. Verified via `Grep "DANDI 000003 breadth probe FAIL; only one IBL-class"` returning a single line and `Grep "DANDI 000003 breadth probe recorded FAIL"` returning zero. No `data-status` flip (already `populated` from 2026-05-07), no marker change (already `STATUS=POPULATED`), no neighbouring-block changes, no CSS / JS / scaffold / page-order / portfolio-taxonomy edits.

**Receipts captured at `/Users/Zer0pa/orchestration-state/receipts/ZPE-Neuro-2026-05-08-wave2/`:**
1. `block-only-1440.png` (1440×760 crop) — slot 17 alone with the trimmed blocker visible: `DANDI 000003 breadth probe FAIL; only one IBL-class target counted. Latency is a modeled cycle figure, not measured on silicon. No claim of …`. SVG renders cleanly with the y-shifted spike-extract / 5-seed replay layout (no caption overlap).
2. `in-context-1440.png` (1440×2300 crop) — three rows of the continuation grid: row above (slot 14 Indus-Valley + slot 15 ZPE-FT), slot 17 row (slot 16 ZPE-IoT + slot 17 ZPE-Neuro), row below (slot 18 ZPE-Taste + slot 19 ZPE-Geo). Card height matches its row peer; grid rhythm holds.
3. `full-1440.png` — 1440×18000 full-page raw capture for archival.

**Receipt NOT captured: mobile (375–414px) screenshot.**
- Two attempts via `Google Chrome --headless --screenshot --window-size=414,N` failed: at N=32000 / 20000 / 14000 the output file was 78–122 KB (Chrome's `tile memory limits exceeded` warning fires; page renders blank past ~8000 px); at N=12000 the file rendered correctly but ended above slot 17.
- Two attempts via `python3.11 playwright` failed because the dev server at `:4631` flapped to HTTP 500 (Miniflare/undici `fetch failed` — Cloudflare-Vite plugin runtime error in shared infrastructure) during the second context creation. Restarting `bun run dev` is out of corridor-agent scope per the brief.
- Mitigation: the mobile fold path is governed by shared CSS I have not touched — `.lanes-grid` collapses to single column at `max-width: 980px`, `.lane-body` collapses to single column at `max-width: 720px` (`public/landing-rd-v3/styles.css:294`). My viz uses the standard `viewBox="0 0 400 260" preserveAspectRatio="none"` `class="svg-soft"` spec, identical to DM3 / Energy / all reference cards. Wave-2 §C.5 classifies substrate vizes (mine) as "generally hold — the visual structure scales to a wider/shorter rectangle without losing readability." Mobile fold inheritance is mechanical; no card-level deviation possible without a CSS edit, which is out of scope.

**Benchmark-footprint and style compliance:**
- Headline: `Spikes you can <span class="accent">replay bit-exact.</span>` — 32 chars, 5 words, accent on last 2 words ending with period. Wraps to 2 lines on the populated grid (consistent with row peer ZPE-IoT "Ten wins on declared ground." and reference ZeroShip "A freighter that ships. Zero CO₂."). Within §C.1 4–7 word / 25–40 char envelope.
- Lane-tag cat: `· ZPE-NEURO · SPIKE EXTRACTION` — 4 words, 30 chars. Within §C.1 ≤38-char envelope.
- Human paragraph: ~342 chars, ~58 words, one `<strong>` span anchoring `DANDI 000034 at 401× event-ratio compression and 78.44 µV RMSE`. Within §C.1 250–400-char envelope.
- Metrics: 4 rows (`401×` / `78.44 µV` / `5 / 5 SEEDS` / `C & D PASS`). Conforms to the §C.2 4-metric mandate.
- Viz: 3-section schematic (raw waveform → extracted spike events → 5-seed bit-exact replay), white-only strokes, three JetBrains-Mono callouts, viewBox 0 0 400 260, viz-cap reserved bottom strip y > 218. Markup ~3.0 KB — within §C.1 800–4500-byte envelope.
- Caption: `FIG · DANDI 000034 · BIT-EXACT SPIKE EXTRACTION` — 47 chars, names what the figure IS, within §C.3 ≤50-char envelope. (Renders on 2 lines at the rendered viz column width — minor display wrap, within envelope.)
- Blocker: 235 chars, 35 words, one paragraph, plain prose, no nested tags besides the leading `.k` span. Within §C.2 ≤280-char ceiling and §C.1 25–40-word envelope.
- CTA: `OPEN REPO SURFACE` → `https://github.com/Zer0pa/ZPE-Neuro` per §B Slot 17. `data-repo="zpe-neuro"`, `target="_blank"`, `rel="noopener"` preserved.
- Scaffold integrity: 4 corner spans on the article, 4 corner spans on the viz, `class="svg-soft"`, `data-repo`, `viz-cap` div all present. No drift from §C.0.

**Repo / web truth-sync performed this Wave-2 pass:** none. Per §B `ZPE-Neuro · ALIGNED · No repo-front-door correction required from this pass.` Repo HEAD remains `3e6d53b`. The earlier-surfaced repo-side asks (G4 GitHub description, G4 topics overrun, G1 visibility/release row, README license-badge mismatch, "Honest Blocker INC" prose drift, stale `commit_sha`) remain surfaced-but-not-landed in the 2026-05-08 prior entry below; Wave-2 §B does NOT re-surface them and explicitly classifies the lane as ALIGNED, so they are not Wave-2 corridor scope. Operator may pick them up under a separate repo-front-door pass when convenient.

**Status:** Wave-2 polish landed. `data-status` remains `populated`. Card prose, metrics, viz, blocker, and CTA are all at reference caliber per §C.4 / §D row 17.

`Lane ZPE-Neuro: wave-2 outcome = fixed (single bounded blocker trim per §C.3 geometry note); repo/front-door status = unchanged (§B = ALIGNED, no repo correction required); landing-card status = populated; receipts ready.`

---

### 2026-05-08 · Landing-card slot 17 POPULATED (Corridor Agent)

**Wave:** Landing-card population on `landing-rd-v3` (per `LANDING_PAGE_CARD_SYSTEM_RESET_2026-05-07.md` + `LANDING_CARD_CENTRALIZED_BRIEF_2026-05-07.md` §A/§B/§C slot 17).

**Block touched:** `<article data-lane="zpe-neuro" data-portfolio="encoding">` and its preceding `<!-- CARD: ZPE-Neuro · ENCODING · STATUS=… -->` marker in `/Users/Zer0pa/Website-fresh/.claude/worktrees/landing-rd-v3/site-v2/src/landing-rd-v3.template.html`. `data-status` flipped `placeholder` → `populated`; comment marker flipped `STATUS=PLACEHOLDER` → `STATUS=POPULATED`.

**Files changed (this session):**
- `Website-fresh/.claude/worktrees/landing-rd-v3/site-v2/src/landing-rd-v3.template.html` — single article block + its CARD comment (one Edit + one atomic Python rewrite for the SVG y-shift).
- No other files in any worktree, repo, or shared scaffold (CSS/JS/page-order/portfolio-taxonomy/neighbouring-card untouched).

**Truth basis used:**
- §A controlling proofs: `reports/ZPE-Neuro.md:24-34,51`; `falsification/ZPE-Neuro.md:35-92`; release v0.1.1; DANDI `000034`.
- §B headline + metric direction: bounded spike-event extraction preview on DANDI 000034; metrics 401× / 78.44 µV / 5/5 seeds / Gate C/D PASS.
- README cross-check (HEAD `3e6d53b`, fetched 2026-05-07 23:29 SAST): metrics rows match the §A envelope verbatim; modeled-not-on-silicon disclaimer carried through; broad-coverage / commercial-readiness implications excluded.

**Card composition:**
- Headline: `Spikes you can <span class="accent">replay bit-exact.</span>` (32 chars; wraps to 2 lines, consistent with row-peer ZPE-IoT and benchmark ZeroShip).
- Lane tag: `· ZPE-NEURO · SPIKE EXTRACTION` (4 words, 30 chars).
- Human paragraph: anchored to DANDI 000034 401× / 78.44 µV win + 5-seed bit-exact replay; latency-modeled-not-measured-on-silicon and breadth-beyond-000034-open disclaimers in-prose.
- Four metrics: Event-ratio compression `401×` / Spike-event RMSE `78.44 µV` / Determinism `5 / 5 SEEDS` / Authority gates `C & D PASS`.
- Bespoke 3-section SVG (`viewBox="0 0 400 260"`, white strokes only, JetBrains Mono callouts): raw extracellular trace → extracted spike-event ticks → 5-seed bit-exact replay column. Y-coordinates re-shifted (via atomic Python rewrite) so y > 218 is reserved for the `.viz-cap` overlay strip; verified in source via `Grep "y=\"170\">S1"` returning 1 hit.
- Caption: `FIG · DANDI 000034 · BIT-EXACT SPIKE EXTRACTION` (47 chars).
- Blocker: `DANDI 000003 breadth probe recorded FAIL; only one IBL-class extracellular target counted as breadth. Latency is a modeled cycle figure, not measured on silicon. No claim of broad neural-decoding coverage or commercial readiness beyond the bounded DANDI 000034 surface.`
- CTA: `OPEN REPO SURFACE` → `https://github.com/Zer0pa/ZPE-Neuro` (per §B; `data-repo="zpe-neuro"`, `target="_blank"`, `rel="noopener"`).

**Diff scope assertion:** every edit is confined to the single `<article data-lane="zpe-neuro">…</article>` span and its immediately preceding `<!-- CARD: -->` comment. Verified by `Grep "data-lane=\"zpe-neuro\""` returning 1 line and visual review of adjacent slots 16 (zpe-iot) and 18 (zpe-taste) whose markers and articles were not touched by this session.

**Receipts captured:**
- `/Users/Zer0pa/orchestration-state/receipts/ZPE-Neuro-2026-05-08/block-only-1440-pre-svg-shift.png` — block-only crop at 1440 viewport, **pre** SVG y-shift; shows headline, paragraph, all four metrics, viz with caption-overlap (subsequently fixed in source).
- `/Users/Zer0pa/orchestration-state/receipts/ZPE-Neuro-2026-05-08/in-context-1440-pre-svg-shift.png` — block-in-context crop at 1440 viewport showing slot 17 paired with row-peer slot 16 (ZPE-IoT) and the rows above/below (placeholders, then GnosisGlyph + Indus-Valley); confirms card height matches the row-peer and the grid rhythm holds.

**Receipts NOT captured (post-SVG-shift, mobile):**
- Post-shift block-only / in-context / mobile screenshots could not be retaken: at resume the `landing-rd-v3` dev server on `:4631` returned HTTP 500 (Miniflare/undici `fetch failed` — Cloudflare-Vite plugin runtime error in shared infrastructure) and then refused TCP connections entirely. Restarting `bun run dev` is out of corridor-agent scope per the brief. SVG fix is verified in the source file via `Grep`. Operator should re-run the dev server and recapture the post-shift screenshots, or fold the recapture into a Website-Orchestrator pass.

**Repo-side asks surfaced (NOT landed by this session):**
- §B Slot 17 lists `G1` + `G4` as needed; §A flags blind-clone status drift and visibility/license doc drift. Concretely on `Zer0pa/ZPE-Neuro` HEAD `3e6d53b`:
  1. **G4 — GitHub description overruns the bounded surface:** currently `ZPE-Neuro V0.0: DETERMINISTIC NEURAL SIGNAL CODEC: Extracellular Spikes | Neuropixels | DANDI-Anchored | IBL Validated | Spike Train Transport`. `V0.0` is stale (PyPI is `0.1.1`) and "DETERMINISTIC NEURAL SIGNAL CODEC" overruns the bounded extracellular spike-event surface.
  2. **G4 — GitHub topics overrun:** `lossless-compression`, `neural-decoding`, `brain-computer-interface`, plus the ECG/EEG-adjacent fan-out. None are evidenced on the bounded DANDI 000034 surface.
  3. **G1 — README visibility / release row stale:** badge text says `release: private staged` while repo + PyPI v0.1.1 are public.
  4. **G2-adjacent — README license badge mismatch:** alt-text says `License: SAL v7.1` but the URL goes to a `SAL%20v7.0` shield.
  5. **README "Honest Blocker" drift:** says `Blind-clone authority replay remains INC` while §A authority + April-24 artifacts say `PASS`.
  6. **A1 (carried from 2026-05-03 Phase-1):** `commit_sha` in `## Repo Identity` is `7d30d52d704e`; HEAD is `3e6d53b`.
- These are listed for operator review only. No PR opened. Do not entangle these with the still-open B-01 license-metadata bundle (pyproject.toml / .zenodo.json / CITATION.cff still on `SAL-7.0` while LICENSE moved to `SAL-7.1` via PR #51) — that bundle remains gated on operator approval per the 2026-05-03 handover.

**Hard "do not" list still honoured:** no sub-agents; no shared CSS/JS/page-order/portfolio-taxonomy/neighbouring-card edits; no `bun run dev` restart; no PyPI `v*` tag push; no B-01 fix PR opened.

**Status:** Slot 17 landing-card POPULATED in source; operator to recapture post-shift screenshots once `:4631` is back. Repo-side G1/G4/blind-clone-drift surfaced for the next gate.

`Lane ZPE-Neuro: repo/front-door status = unchanged (G1/G4 surfaced, no PR opened); landing-card status = populated; receipts ready.`

---

### 2026-05-03 · Phase-1-complete (Lane Agent)

**State report:** `/tmp/lane-ZPE-Neuro/STATE.md`

**Severity summary:** No CRITICAL drift — `audit-readme-fidelity` PASS (18/18 encoding lanes, 7 invariants), live page `200 OK`. Two LOW website-side edits pending operator gate: (A2) extend `description` frontmatter to include the "Install from PyPI" sentence from README para 1; (A3) expand `"INC"` → `"INCONCLUSIVE"` in `honest_blocker` for public readability. One REPO-SIDE MEDIUM flag: `commit_sha` in `## Repo Identity` is stale (`7d30d52d704e`) — actual HEAD is `3e6d53b` (pushed 2026-05-03); repo team must refresh README first, then .md follows. One repo flag for the WTI word-count shortfall (~20 words vs 25-word target, RF-09 universal).

**Status:** Phase 1 complete. Awaiting operator gate before Phase 2/3.

---

### 2026-05-03 · Phase-1 first run (operator) — INITIAL

Source: Initial dispatch. No prior STATE report exists for this lane.

**Ask:** Execute Phase 1 of the canonical brief workflow (see §6 Phase 1 — STATE OF THE LANE in `/Users/Zer0pa/orchestration-state/LANE_AGENT_BRIEF_2026-05-03.md`).

**Output expected** at `/tmp/lane-ZPE-Neuro/STATE.md`:
- GitHub status (visibility / last-commit SHA / README compliance per §3.3 of brief)
- Hugging Face status (`huggingface.co/Zer0pa/ZPE-Neuro` and any related namespaces)
- Website status (build / page render at http://127.0.0.1:4500/encoding/ZPE-Neuro/ / `audit-readme-fidelity` if encoding lane)
- Diff (README vs `src/content/lanes/encoding/ZPE-Neuro.md`, severity-tagged: CRITICAL / MEDIUM / LOW / REPO-SIDE)
- Nuance audit (verdict / boundary list / verification table / over+under-claim)
- Recommended actions (small, ordered, severity-tagged)
- Open questions for operator

Append a `Phase-1-complete` entry above the divider with the path to your STATE.md and a 2-3 sentence severity summary (any CRITICAL drift? Any RF for repo team?).

Do NOT proceed to Phase 2/3 without the operator's gate.

Sonnet sub-agents only if you dispatch any.

**Status:** Awaiting Lane Agent execution.

---

*Lane Agent: read this file end-to-end, then execute the topmost unfulfilled entry. Append your reply with receipts as a new dated entry above the divider when you complete each ask.*