---
vault_clearance: APOLLYON-THAUMIEL
halo:
  classification: OPERATIONAL
  confidence: "SCANNING ACTIVE. Three probes. One shield. Nothing native. Find her home."
  front: "32_Project_HomeWorlds"
  custodian: "Jixiang Leng"
  created: 2026-04-16
  updated: 2026-04-03
  wing: UNASSESSED
  containment: "The probes are real. The deliveries are in the fossil record. The spectral signatures are in the archives. The voids are mapped. The question is not whether they exist but where they are now."
---

# 32_Project_HomeWorlds

> *"Where are they parked?"*

## Purpose

Locate the Von Neumann probes. Three self-replicating machines seeded Earth's biosphere at known epochs. Their delivery vehicles are in the meteorite record. Their spectral fingerprints are in the ISM. Their territories are visible as cosmic voids. This project builds the spatial scanning matrix to find their current positions, operating states, and territory boundaries.

**Public surface:** Spatial scanning matrix for anomalous ISM signatures using NASA/ESA archival data.
**Private surface:** Probe triangulation. Where is she. Where are the others. What is the Moon doing about it.

### How this fits the argument

> **In one sentence:** HomeWorlds scans the interstellar medium for von Neumann probe signatures — extending the vault's measurement framework from cells to cosmic spatial structure.

**The vault's unified argument** is in [THE_ARGUMENT.md](../THE_ARGUMENT.md). This project provides **cosmological context — ISM scanner, spatial scanning, probe detection**.

**Read first:** this README | **Depends on:** [29 Deathomatica](../29_Project_Deathomatica/README.md) (orbital) | **Feeds into:** [05 LENG](../05_Project_LENG/README.md) (constraints)

**Triad:** [BOUNTY_BOARD.md](BOUNTY_BOARD.md) · [WORLDLINE.md](WORLDLINE.md) · [FORM.md](FORM.md) · [BOOK.md](BOOK.md) (vault [FORM_PROTOCOL.md](../FORM_PROTOCOL.md) sections A–H + [TRUTH_PROTOCOL.md](../TRUTH_PROTOCOL.md); on-disk map in FORM appendix; BOOK = STARS / datasets / software per [BOOK_Protocol.md](../BOOK_Protocol.md)) · **Ring #32 of 33** — [`_VAULT_STATE.md`](../_VAULT_STATE.md).

## Full stack applicability (design point)

HomeWorlds is written so the **same vault spine** (FORM / BOOK / WORLDLINE / bounties / tests) applies across goals: **ephemeris & charts**, **large-survey structure** (voids, catalogs, CMB), **transients**, **minor-planet / MPC-class work**, and **exoplanet / mission-archive statistics** are all **in scope as instruments** — different scripts and datasets, one custody story. That is the **point** of stacking them here: not that every pipeline is finished in one file, but that nothing in the design *blocks* you from adding the next instrument without a new project philosophy.

Canonical theory (instruments + ephemeris): [theory/ASTRO_INSTRUMENTS.md](theory/ASTRO_INSTRUMENTS.md). **Limits = bounties:** [BOUNTY_BOARD.md](BOUNTY_BOARD.md) (HW-15+).

## Astrological modeling (public toolchain)

Traditional **tropical** chart geometry (bodies, major aspects, optional Placidus houses) lives in **`scripts/celestial_chart_pipeline.py`**, documented in [theory/ASTRO_INSTRUMENTS.md](theory/ASTRO_INSTRUMENTS.md) §2.

| Tier | Install | When to use |
|------|---------|-------------|
| **Swiss** (`--engine swisseph`) | [public/requirements-astrology.txt](public/requirements-astrology.txt) | Full houses, Asc/MC, nodes, Lilith, Chiron — **industry-standard** Swiss Ephemeris math. |
| **Portable** (`--engine portable`) | [public/requirements-astrology-portable.txt](public/requirements-astrology-portable.txt) | **Skyfield + JPL DE421** (public kernel); geocentric tropical longitudes + aspects **without** houses — works when `pyswisseph` has no wheel (e.g. Windows + Python 3.12 without MSVC). |
| **Auto** | install either or both | Tries Swiss first, then portable. |

**Export JSON / Markdown:** pass `--json-out` and/or `--markdown-out` (ISO-8601 `--datetime`, `--lat`, `--lon`, `--house`). Example: `--engine portable --datetime 2024-03-20T15:00:00+00:00 --lat 31.7683 --lon 35.2137 --json-out data/astrology_runs/chart.json` (folder is gitignored by default).

**Regression suite:** `tests/fixtures/golden_portable_charts.json` locks tropical **longitudes and signs** (Sun–Pluto) for four UTC moments (Greenwich J2000, Tokyo equinox snapshot, Tel Aviv solstice snapshot, Jerusalem spring). `pytest` compares live DE421 output to that file; refresh with `scripts/export_golden_chart_fixtures.py`. International observatory and ephemeris anchors (East Asia, Israel, IAU) are listed in [BOOK.md](BOOK.md).

**Cross-engine merge:** `scripts/ephemeris_hub.py` combines **DE421 portable**, **JPL Horizons** (observer ecliptic Q31, geocentric), and **optional Swiss** in one JSON bundle. The chart CLI accepts **`--compare-horizons`** to attach Horizons deltas to `--json-out` (network required).

**Planetary / sign design tests:** `tests/test_planetary_detection_suite.py` checks that all ten major bodies are present with finite tropical longitudes, signs match 30° sectors, aspects fire on synthetic geometry, geocentric invariance holds, golden seasonal Sun stories hold, and (opt-in) live Horizons + Ceres (`HOMEWORLDS_NETWORK_TESTS=1`). This validates **ephemeris reproduction**, not blind survey discovery.

This lane is **orthogonal** to the ISM / void scanner: same vault project, different instrument — limits as **OPEN bounties:** [BOUNTY_BOARD.md](BOUNTY_BOARD.md) HW-15+.

### Windows PowerShell (paths with `[` `]`)

If the vault folder name includes brackets (date-prefixed Obsidian vault), plain `cd` can fail. **Dot-source** [`scripts/Enter-HomeWorlds.ps1`](scripts/Enter-HomeWorlds.ps1) (leading dot required) so **`PYTHONPATH` stays in this window**. Easiest one-liner from **any** cwd — use a **single-quoted** full path so brackets are literal:

```powershell
. 'C:\Users\you\...\32_Project_HomeWorlds\scripts\Enter-HomeWorlds.ps1'
```

Or move to `scripts` first with **`Set-Location -LiteralPath`**, then:

```powershell
Set-Location -LiteralPath 'C:\Users\you\...\32_Project_HomeWorlds\scripts'
. .\Enter-HomeWorlds.ps1
```

That sets **`PYTHONPATH`** to this project’s `scripts` directory so `python -m sky4d.build`, `pytest`, and `import sky4d` work. One-time deps from project root (after the script runs, your cwd is `32_Project_HomeWorlds`): `pip install -r public/requirements.txt -r requirements-dev.txt`.

If the terminal is already at the **vault root** (workspace root), you can instead run:

```powershell
$hw = Join-Path (Get-Location) '32_Project_HomeWorlds'
Set-Location -LiteralPath $hw
$env:PYTHONPATH = Join-Path (Get-Location) 'scripts'
```

## The Three Probes

| Probe | Name | Payload | Arrival | Spectral Signature | Territory |
|-------|------|---------|---------|-------------------|-----------|
| P1 | THE CODE | Viruses / RNA / DNA | 4.0-3.8 Gya (LHB) | Nucleobase UV absorption (260 nm) | Every genome |
| P2 | THE LIGHT | Cyanobacteria / Photosynthesis | ~3.5 Gya | Porphyrin Soret band (400 nm), Q-bands (500-650 nm) | Oxygen atmosphere |
| P3 | THE RECYCLER | Fungi / Melanin / Spore | ~3.5-2.0 Gya (continuous) | Broadband UV (peak ~210 nm), 2175 A bump, AME (20-30 GHz), ARCADE excess (22 MHz-10 GHz) | Cosmic voids, every lysosome |

**Shield:** The Moon. Quarter-phase period (7.38 d) jams Ganymede's broadcast (7.155 d) via Z3/IDO1/tryptophan depletion. Beat frequency: 234.6 days.

## Origin Chain (HomeWorlds / StarSight)

> *Where did she come from?*

### What's measured (no modeled dates):

| Evidence | Tier | What it is |
|----------|------|------------|
| Pre-Cambrian fungal-like fossils are simple filaments | P | Morphology, SEM |
| Post-Cambrian fungi have full decomposition toolkit | P | Morphology discontinuity |
| The jump happens at/near the Cambrian boundary | P | Stratigraphic position |
| Acraman impact ejecta layer exists (shocked minerals) | S | Mineral ID, shock features |
| Biological turnover immediately ABOVE Acraman ejecta | P | Acritarch species counts |
| MAPCIS (~600 km, Australia) has lonsdaleite + iridium | S | Mineral/elemental detection |
| MAPCIS sits at Precambrian-Cambrian boundary | P | Stratigraphic position |
| Gardnos crater (Norway, 5 km) at 546 Ma (U-Pb zircon) | S/D | Best-dated boundary-era crater |
| Gardnos has Cambrian fossils in post-impact fill | P | Fossil in fill = crater predates Cambrian |
| Lysosome machinery was complete at LECA | S | Genomic — universal in all eukaryotes |
| BLOC-1/2/3 sorting already bifurcated before LECA | S | Phylogenetic analysis (PNAS 2024) |
| IOM in meteorites spectrally matches allomelanin | S | FTIR, Raman |
| Murchison meteorite contains nucleobases | S | Mass spec |

### The chain (stratigraphic order, no absolute dates):

```
Simple filaments in caves (below Cambrian)
    ↓
IMPACT: Acraman (ejecta layer, biological turnover above)
    ↓
IMPACT: Gardnos (crater with Cambrian fossils in fill)
    ↓
IMPACT: MAPCIS? (unconfirmed 600 km structure at boundary)
    ↓
CAMBRIAN BOUNDARY
    ↓
Full-stack fungi appear (decomposition toolkit, mycorrhizae, melanin arsenal)
    ↓
Lysosome already present in ALL eukaryotes (she's inside every cell)
```

### What StarSight does with this:

Acraman and Gardnos have measured CRATER GEOMETRIES (ellipticity, ejecta asymmetry) that constrain the IMPACT VECTOR. From the vector: back-propagate an orbit (deterministic, Kepler + perturbations). From the orbit: constrain the source body (must support organic survival, must be carbonaceous, must have melanin-compatible chemistry).

The constraint graph maps every link. The Daemon scores its own completeness (comfort = 0.685 on the void network). Applied to the origin chain, it will tell us how much of the picture we can see and WHERE THE BOUNDARY IS — what measurement we're missing.

### Open questions:

1. What is the impact vector for Acraman? (Published: crater asymmetry + ejecta distribution)
2. What is the MAPCIS impactor composition? (Lonsdaleite = carbonaceous? Or iron?)
3. Does IOM in Acraman-associated sediments match fungal melanin spectrally?
4. What C-type asteroids have orbits that could deliver to Earth at the right stratigraphic position?

See: [theory/HALO_THE_FOG_OF_THE_PAST.md](theory/HALO_THE_FOG_OF_THE_PAST.md) for full audit of dating assumptions.

## What We're Scanning For

### Probe 3 (The Recycler / Lady of Graves)

She has FOUR spectral bands, all with unknown or unconfirmed carriers:

| Band | Wavelength/Frequency | Feature | Carrier Status | Years Unsolved |
|------|---------------------|---------|----------------|----------------|
| UV | 217.5 nm | 2175 A extinction bump | UNKNOWN | 60+ |
| Visible | 400-900 nm | 500 DIBs | UNKNOWN (only C60+ confirmed) | 100+ |
| Microwave | 10-60 GHz (peak ~25 GHz) | AME / spinning dust | UNCONFIRMED | 25+ |
| Radio | 22 MHz - 10 GHz | ARCADE excess / Space Roar | UNKNOWN | 17+ |

One material. Four bands. All pointing at carbonaceous organic macromolecules. The public literature has been looking at these separately for decades. Nobody has overlaid them on a spatial map and asked: do they come from the same places?

### The Noise They Threw Away

Every CMB experiment subtracts "foreground contamination." That contamination IS the data:

| Foreground | Frequency | What It Actually Is |
|-----------|-----------|-------------------|
| Synchrotron | 1-30 GHz | Relativistic electrons in magnetic fields (IDENTIFIED) |
| Free-free | 1-100 GHz | Thermal electron-ion scattering (IDENTIFIED) |
| AME / spinning dust | 10-60 GHz | "Spinning PAH nanograins" (UNCONFIRMED) |
| Thermal dust | 100-1000 GHz | Silicate + carbonaceous grains (partially identified) |
| ARCADE excess | 22 MHz-10 GHz | 6x brighter than all known sources (UNKNOWN) |

The ARCADE excess is the smoking gun. Power law with spectral index -2.56, isotropic, 6x above theory. Nobody can explain it. It looks exactly like what a universe-spanning broadband absorber/re-emitter would produce.

### Anomalies That Survive Subtraction

| Anomaly | Location | Temperature | Status |
|---------|----------|-------------|--------|
| CMB Cold Spot | RA 3h15m, Dec -19.6 deg (Eridanus) | -70 to -140 uK | UNEXPLAINED |
| Axis of Evil | Aligned with ecliptic | Low multipole alignment | UNEXPLAINED |
| Hemispherical asymmetry | Full sky | One hemisphere cooler | UNEXPLAINED |
| Cosmic dipole excess | (l,b) = (264, 48) | Matter dipole 6.4 sigma too large | UNEXPLAINED |

## Scanning Architecture

### Data Sources (all public, all archival, all executable NOW)

| Source | What It Gives Us | Access |
|--------|-----------------|--------|
| SDSS (DR17+) | Galaxy positions, spectra, void catalogs | CasJobs SQL / API |
| Pantheon+ | 1701 SNe Ia with multi-band photometry | GitHub CSV |
| Planck (PR3/PR4) | CMB, dust, AME, synchrotron maps (HEALPix FITS) | IRSA / ESA archive |
| WMAP | 9-year foreground maps | LAMBDA |
| CHIME/FRB | Fast radio burst catalog with DM, positions | CHIME website |
| JWST archive | UV/NIR spectroscopy at z=2-7+ | MAST |
| HST UV | UV spectroscopy of background quasars | MAST |
| IRSA Dust Service | E(B-V) reddening at any coordinate | API |
| VIDE | ~1500 cosmic voids from SDSS (ZOBOV) | cosmicvoids.net |
| JPL Horizons | Solar system ephemerides, any body | API |

### Spatial Scanning Matrix

The engine cross-matches these databases against probe territory predictions:

```
For each sightline (RA, Dec, z):
  1. Classify: void / filament / wall (from VIDE or SDSS void catalog)
  2. Query: E(B-V) reddening (IRSA)
  3. Query: 2175 A bump strength (if UV spectrum exists in HST/JWST)
  4. Query: AME signal (Planck pixel at that position)
  5. Query: dust optical depth (Planck tau map)
  6. Query: FRB DM residual (if CHIME source at that position)
  7. Compute: excess = observed - (synchrotron + free-free + thermal dust)
  8. Flag: does the excess follow the melanin absorption/emission profile?
```

### The 10 Archival Tests (from HALO_THE_SIGNAL_MAP)

| # | Test | Key Question |
|---|------|-------------|
| 1 | 2175 A bump: void vs filament sightlines | Is the absorber concentrated in voids? |
| 2 | H0 by direction relative to KBC Void | Does absorption vary with void geometry? |
| 3 | H0 from z>0.1 only vs Planck | Does the bias disappear outside the void? |
| 4 | AME in Bootes Void | Is there spinning dust where nobody looked? |
| 5 | Void galaxy SFR vs controls | Is star formation actively suppressed? |
| 6 | FRB DM residuals: void vs filament | Are electrons locked in IOM grains? |
| 7 | 2175 A at z=2-7 (JWST) | Does the absorber accumulate over cosmic time? |
| 8 | Void boundary sharpness vs LCDM | Are edges sharper than gravity predicts? |
| 9 | Cosmic dipole vs void geometry | Does the flow follow the fairy ring? |
| 10 | Wavelength-dependent H0 | Does tension track the melanin spectrum? |

## Aggregated Spatial Tools (Vault Inventory)

### From 21_Project_LuaOversoul

| Script | Capability |
|--------|-----------|
| `moon_deep_time_probe.py` | JPL Horizons API, REBOUND backward integration |
| `gu_moon_mass_hypothesis.py` | Horizons state vectors, REBOUND N-body comparison |
| `siderophile_strata_monte_carlo.py` | Monte Carlo geochemistry constraints |

**Documented but not scripted:** SPICE/SpiceyPy, Skyfield, REBOUNDx, JPL SBDB, MPC.

### From 29_Project_Deathomatica

| Script | Capability |
|--------|-----------|
| `deathomatica_engine.py` | Full N-body + tidal evolution, 4.5 Gyr backward |
| `jovian_tidal_analytical.py` | Analytical tidal ODE, Laplace resonance |
| `ganymede_phase_lock.py` | Phase-lock test, beat frequency analysis |
| `constraint_engine.py` | LENG cosmological residuals, absorber physics |
| `constraint_engine_live.py` | IRSA dust API, Pantheon+ fetch, chromatic H0 |
| `void_spectral_graph.py` | Void network topology, graph Laplacian |
| `void_absorption_atlas.py` | ISM absorption catalog, Hodge-de Rham, persistence |

### From 28_Project_RedFromTheGrave

| Asset | Capability |
|-------|-----------|
| `grave_atlas/globe/` | Three.js 3D WGS84 globe with territory overlays |
| Territory PNGs | Fungal territory + cancer atlas geographic layers |

### From 30_Project_Crucible

| Script | Capability |
|--------|-----------|
| `spectral_k_analysis.py` | Coupling tensor eigenvalue spectrum |

### From 05_Project_LENG

| Module | Capability |
|--------|-----------|
| `lotus/cosmology.py` | All LENG cosmological predictions (zero parameters) |
| `lotus/core/geometry.py` | S5/Z3 invariants |

## Avoiding the Orthodox Traps

### What We Do NOT Do

| Orthodox Trap | Why It Fails | What We Do Instead |
|--------------|-------------|-------------------|
| Chi-squared fitting | Introduces free parameters; lets you fit an elephant | Arithmetic residuals: measured - predicted. The residual IS the data. |
| Bayesian priors | Prior choice determines posterior; unfalsifiable | Zero priors. LENG predictions have zero free parameters. |
| Monte Carlo model selection | Compares models, not reality | One model (LENG geometry) vs one measurement. The gap is the signal. |
| "Consistent with random" | Everything is consistent with random at low N | Don't compare to random. Compare to LENG prediction. The deviation is the map. |
| kNN with arbitrary k | k is a free parameter masquerading as a "choice" | Vietoris-Rips complex: one scale parameter (epsilon) with physical meaning |
| Ensemble comparison | Choosing the ensemble is the assumption | No ensemble. Persistence over scale range: features that survive are real. |
| "Statistically significant" | p-values answer the wrong question | Does the spatial pattern match the spectral pattern? Binary. |

### The Practical Questions

1. **Where is the 2175 A bump strongest?** Map it. Plot it. Does it trace void boundaries?
2. **Where is AME strongest outside the galactic plane?** Nobody has made this map for voids.
3. **Does the ARCADE excess have spatial structure?** If it's isotropic, that's a claim. Verify it at the resolution Planck gives us.
4. **Do void sightlines show more reddening than filament sightlines at matched redshift?** One query. One answer.
5. **Is the chromatic H0 gradient real?** We plotted it (Deathomatica). Now test it with Pantheon+ multi-band.
6. **Do FRB dispersion measures drop in voids?** If electrons are locked in IOM grains, yes.
7. **Are void walls sharper than LCDM predicts?** One measurement. Fairy ring or gravity?

## WORD Protocol

**W**orking **O**bjects = **R**ecording = **D**istribution.

Every instrument carries its bias genealogy WITH it. The JSON IS the data AND the assumptions AND the classification tier. Nothing ships without its chain of custody from photon to catalog value. Tools are dead. People are alive. Information is immortal.

See [theory/HALO_THE_SCANNER.md](theory/HALO_THE_SCANNER.md) for the full bias audit of every dataset we touch.

## Project Structure

```
32_Project_HomeWorlds/
+-- README.md                              <- This file (PRIVATE)
+-- BOUNTY_BOARD.md                        <- Open bounties + solved HW ids
+-- WORLDLINE.md                           <- Run log / narrative
+-- FORM.md                                <- FORM_PROTOCOL A–H + TRUTH axis; appendix = paths
|
+-- public/
|   +-- README.md                         <- PUBLIC surface for GitHub release
|   +-- requirements.txt                  <- Python dependencies
|   +-- .gitignore                        <- Git ignore rules
|
+-- theory/
|   +-- HALO_THE_SCANNER.md               <- TRUTH FORM: bias audit + measurement genealogy
|
+-- scripts/
|   +-- instruments.py                    <- Living instrument framework (6 instruments)
|   +-- homeworlds_scanner.py             <- Spatial scanning engine (IRSA + 37 targets)
|   +-- run_tests.py                      <- Archival tests with real Pantheon+ data
|   +-- negative_space_scanner.py         <- Exclusion zone topology + Dirac operator
|   +-- vast_parser.py                    <- VAST void catalog parser (ZOBOV + VoidFinder)
|
+-- data/
|   +-- pantheon_plus.dat                 <- 1701 SNe Ia from Pantheon+SH0ES (real data)
|   +-- homeworlds_master_v2.json         <- All instruments composed
|   +-- archival_tests_v1.json            <- Test 2+3 results with real data
|   +-- negative_space_v1.json            <- v1 scan (29 hardcoded voids)
|   +-- negative_space_v2_vast.json       <- v2 scan (217 VAST VIDE interior voids)
|   +-- vast_vide_summary.json            <- VAST VIDE catalog summary
|   +-- vast_voidfinder_summary.json      <- VAST VoidFinder catalog summary
|   +-- vast_catalog/
|   |   +-- V2_VIDE-*_Planck2018_*.dat   <- 531 ZOBOV voids (Douglass+2023)
|   |   +-- VoidFinder-*_Planck2018_*.txt <- 1163 maximal spheres + 39736 holes
|   +-- figures/
|   |   +-- pantheon_hubble_kbc.png       <- Hubble diagram colored by KBC membership
|   |   +-- pantheon_directional_sky.png  <- Hubble residuals on sky (Mollweide)
|   |   +-- homeworlds_sky_map.png        <- All 37 scan targets
|   |   +-- homeworlds_spectral_fingerprint.png <- Four anomalous spectral bands
|   |   +-- homeworlds_chromatic_h0.png   <- Published H0 vs melanin prediction
|   |   +-- homeworlds_ebv_by_category.png <- Dust reddening by target category
|   |   +-- negative_space_sky.png        <- Exclusion map (Mollweide)
|   |   +-- negative_space_persistence.png <- Betti curves vs random baseline
|   |   +-- negative_space_dirac.png      <- Dirac spectrum + chirality
```

## Dependencies

- **29_Project_Deathomatica:** Constraint engine, void absorption atlas, all theory HALOs
- **28_Project_RedFromTheGrave:** Globe visualization, territory overlays, Ganymede hypothesis
- **21_Project_LuaOversoul:** Orbital mechanics engine, Horizons API, REBOUND
- **30_Project_Crucible:** Coupling tensor, spectral analysis
- **05_Project_LENG:** Zero-parameter cosmological predictions

## Bounty Board

| ID | Task | Status |
|----|------|--------|
| HW-1 | Build spatial scanning engine (aggregate vault tools) | **DONE** |
| HW-2 | VIDE void catalog download + parser | **DONE** — VAST/Zenodo, 531 VIDE + 1163 VoidFinder |
| HW-3 | Planck foreground map pixel extraction (AME, dust, synchrotron at void positions) | OPEN |
| HW-4 | SDSS CasJobs query: void galaxy SFR vs field controls | OPEN |
| HW-5 | HST UV archival query: 2175 A bump along void sightlines | OPEN |
| HW-6 | CHIME FRB catalog: DM residuals by void membership | OPEN |
| HW-7 | Pantheon+ directional H0 analysis (KBC geometry) | **DONE** — Tests 2+3 with 1701 SNe |
| HW-8 | ARCADE excess spatial structure verification | OPEN |
| HW-9 | Probe triangulation map (overlay all 4 spectral bands) | OPEN |
| HW-10 | Public release: spatial scanning matrix (NASA data only) | **READY** — public/ structure built |
| HW-11 | Negative space scanner v1 (exclusion zones + persistent homology) | **DONE** — 29 voids |
| HW-12 | Negative space scanner v2 (VAST catalog, 217 interior voids) | **DONE** — beta_1 +5.8 sigma |
| HW-13 | Dirac operator on void network (daemon v3 pattern) | **DONE** — eta=+0.0024, 94 sinks / 123 sources |
| HW-14 | Public polish: pins, CI smoke, narrative sync with vault triad | OPEN |

---

*Created April 16, 2026. Three probes built this biosphere. One of them is thinning the wall. Her spectral signature is in four wavelength bands, all unsolved, all pointing at the same material. The archives are full of her fingerprints. Nobody looked because nobody was looking. We are looking now.*


<!-- live-edge-test 13:47:36 -->