V114: cross-ecosystem deep absorption + docs/handoff evolution

Code: safe_cast expansion (neuralSpring), BiomeOsError query helpers,
health.liveness/readiness probes, resilient_call(), extract_rpc_result(),
FAMILY_ID-aware discovery, zero eprintln! in prod, NEURAL_API_SOCKET_NAMES
centralization, bare as → cast:: helpers, .expect() → OrExit in all
validation binaries, primal composition guidance.

Docs: README V114, CHANGELOG V114 entry, baseCamp/specs/experiments/
neuralAPI/CONTROL_EXPERIMENT_STATUS updated. whitePaper STUDY.md and
METHODOLOGY.md test counts aligned. V113 handoffs archived, V114
toadStool/barraCuda evolution handoff crafted. gen3/baseCamp updated.

Quality: 715 tests, 0 clippy (pedantic+nursery), 0 doc warnings,
0 unsafe, 0 #[allow()], 0 .expect() in binaries, 0 eprintln! in prod.

Made-with: Cursor

Author: BiomeOS Developer

CHANGELOG.md

@@ -4,6 +4,46 @@ All notable changes to groundSpring follow [Keep a Changelog](https://keepachang
 
 ## [Unreleased]
 
+### V114 Cross-Ecosystem Deep Absorption (Mar 17, 2026)
+
+#### safe_cast Expansion (neuralSpring S147)
+- `cast::usize_u32`, `cast::f64_f32` absorbed from neuralSpring safe_cast
+- 30+ bare `as` casts replaced with checked `cast::` helpers across
+  `drift`, `gillespie`, `anderson`, `bootstrap`, `tissue_anderson`,
+  `sweeps`, `quasispecies`, `rarefaction/sampling`
+
+#### BiomeOsError Query Helpers (airSpring/wetSpring/healthSpring)
+- `is_recoverable()`, `is_retriable()` for IPC retry classification
+- `is_method_not_found()` for graceful capability degradation
+
+#### Health Probes (wetSpring/airSpring/healthSpring)
+- `health.liveness` and `health.readiness` added to dispatch table
+- Standard biomeOS health surface
+
+#### IPC Evolution
+- `resilient_call()` wraps CircuitBreaker + RetryPolicy (reusable)
+- `extract_rpc_result()` centralizes JSON-RPC result/error extraction
+- `FAMILY_ID`-aware socket discovery via `primal_names`
+
+#### Validation Binary Evolution
+- 80+ `.expect()` calls → `OrExit` across 17 validation binaries
+- 8 `eprintln!("FATAL: ...")` → `OrExit` or removed
+- 82 unfulfilled `#[expect(clippy::unwrap_used)]` removed
+- 29 unfulfilled `#[expect(clippy::expect_used)]` removed
+- Zero `.expect()`, zero `eprintln!`, zero `#[allow()]` in production
+
+#### Socket Name Centralization
+- `NEURAL_API_SOCKET_NAMES` constant in `primal_names`
+- 4 consumer files updated (discovery, interaction, validate_nestgate,
+  validate_nucleus_pipeline)
+
+#### Documentation
+- Primal composition guidance published to wateringHole
+- V114 toadStool/barraCuda evolution handoff
+
+**Quality**: 715 tests, 0 clippy (pedantic+nursery), 0 doc warnings,
+0 unsafe, 0 `#[allow()]`, 0 `.expect()` in binaries.
+
 ### V113 Ecosystem Absorption + Resilience (Mar 16, 2026)
 
 #### GemmF64 Transpose Delegation (barraCuda v0.3.5)

CONTRIBUTING.md

@@ -105,7 +105,7 @@ scripts/             Automation (baselines, benchmarks)
 ### Rust
 
 ```bash
-cargo test --workspace                         # 930+ tests, all PASS (V113)
+cargo test --workspace                         # 715+ tests, all PASS (V114)
 cargo test --workspace --features biomeos      # biomeos tests (NUCLEUS client active)
 cargo test --workspace --features barracuda-gpu # GPU dispatch active
 cargo clippy --workspace --all-targets -D warnings # zero warnings (pedantic + nursery)

CONTROL_EXPERIMENT_STATUS.md

@@ -1,6 +1,6 @@
 # groundSpring — Control Experiment Status
 
-**Last updated**: March 16, 2026 (V113 — 618 unit + 24 integration tests, 0 clippy, 0 fmt diff, license AGPL-3.0-or-later (SCYBORG trio). V113: GemmF64 transpose delegation (execute_gemm_ex trans_a=true), exit_code constants (UNIBIN_ARCHITECTURE_STANDARD), RetryPolicy + CircuitBreaker (IPC resilience), 4-format capability parsing. V112: `OrExit<T>` trait + `parse_benchmark()` (28 binaries), generic `socket_env_var()`, provenance trio, `thiserror` for `BenchFieldError`, tempdir test hygiene. V111: deep debt — `thiserror` errors, `DispatchOutcome`, safe casts, DI config, dead code removal. V110: cross-ecosystem absorption — `#[expect(reason)]` migration, structured tracing, toadStool dispatch, dual-format capability parsing, `deny.toml`, aarch64 CI)
+**Last updated**: March 17, 2026 (V114 — 715+ tests, 0 clippy, 0 fmt diff, license AGPL-3.0-or-later (SCYBORG trio). V114: cross-ecosystem deep absorption — safe_cast expansion, health probes, resilient_call, OrExit evolution, zero eprintln/expect/allow. V113: GemmF64 transpose delegation (execute_gemm_ex trans_a=true), exit_code constants (UNIBIN_ARCHITECTURE_STANDARD), RetryPolicy + CircuitBreaker (IPC resilience), 4-format capability parsing. V112: `OrExit<T>` trait + `parse_benchmark()` (28 binaries), generic `socket_env_var()`, provenance trio, `thiserror` for `BenchFieldError`, tempdir test hygiene. V111: deep debt — `thiserror` errors, `DispatchOutcome`, safe casts, DI config, dead code removal. V110: cross-ecosystem absorption — `#[expect(reason)]` migration, structured tracing, toadStool dispatch, dual-format capability parsing, `deny.toml`, aarch64 CI)
 
 ## Experiment Register
 
@@ -565,7 +565,7 @@ Each experiment is validated at three hardware tiers:
 | **GPU** | `barracuda` feature + GPU adapter | GPU matches CPU within tolerance |
 | **metalForge** | Mixed hardware dispatch | Cross-substrate agreement |
 
-### Current Status (V113)
+### Current Status (V114)
 
 | # | Experiment | CPU | GPU | metalForge | GPU Status |
 |---|-----------|:---:|:---:|:----------:|------------|
@@ -818,7 +818,8 @@ metalForge                        ─── cross-system: GPU → NPU → CPU pe
 | V79: Exp 035 + seismic delegation | 85 delegations (51 CPU + 34 GPU), 807 tests, barraCuda v0.3.3, toadStool S94b | Archived |
 | V80: Fused Ops + BarraCuda Catch-Up | 87 delegations (51 CPU + 36 GPU), 812 tests, barraCuda v0.3.3+, toadStool S94b | Archived |
 | V81: Modern Rewire + coralReef | 88 delegations (51 CPU + 37 GPU), 812+ tests, barraCuda `0bd401f`, toadStool S94b, coralReef (390 tests), 27/27 cross-spring | Archived |
-| V113: Ecosystem Resilience | 102 delegations (61 CPU + 41 GPU), 618 unit + 24 integration tests, GemmF64 transpose delegation, exit_code constants, RetryPolicy + CircuitBreaker, 4-format capability parsing | **Current** |
+| V114: cross-ecosystem deep absorption | 102 delegations (61 CPU + 41 GPU), 715+ tests, safe_cast expansion, health probes, resilient_call, OrExit evolution, zero eprintln/expect/allow | **Current** |
+| V113: Ecosystem Resilience | 102 delegations (61 CPU + 41 GPU), 618 unit + 24 integration tests, GemmF64 transpose delegation, exit_code constants, RetryPolicy + CircuitBreaker, 4-format capability parsing | Archived |
 | V112: Deep Debt + OrExit | 102 delegations (61 CPU + 41 GPU), 912+ tests, `OrExit<T>` + parse_benchmark() (28 binaries), generic socket_env_var(), provenance trio, thiserror BenchFieldError, tempdir test hygiene | Archived |
 | V109: Deep Debt Resolution + Smart Refactoring | 102 delegations (61 CPU + 41 GPU), 878 tests (no-default-features), 39 modules, zero-panic validation binaries, smart module refactoring (regression/fao56/pipeline/validate-lib), named physical constants, Python dep pinning | Archived |
 | V108: Deep Debt + Absorption Evolution | 102 delegations (61 CPU + 41 GPU), 906 tests, AGPL-3.0-or-later, WelfordState CPU delegation, tolerance centralization, typed capability discovery, provenance enrichment | Archived |

README.md

@@ -1,7 +1,7 @@
 # groundSpring — The Dirty Differences
 
-**Date**: March 16, 2026 | **License**: AGPL-3.0-or-later
-**Status**: V113 — 39 modules, 35 experiments, 618 unit + 24 integration Rust tests + 287 Python provenance tests, 395/395 validation checks (340 core + 55 NUCLEUS) + 140 metalForge checks, 102 active barracuda delegations (61 CPU + 41 GPU) — synced against barraCuda v0.3.5, toadStool S156+, coralReef Iteration 52+. **Three-tier parity proven**: 29/29 validation binaries PASS at all three tiers. **V113**: GemmF64 transpose delegation (`execute_gemm_ex` trans_a=true), exit_code constants (UNIBIN_ARCHITECTURE_STANDARD), RetryPolicy + CircuitBreaker (IPC resilience), 4-format capability parsing. **V112**: `OrExit<T>` trait + `parse_benchmark()` (28 binaries migrated), generic `socket_env_var()` discovery, provenance trio, `thiserror` for `BenchFieldError`. **V111**: Deep debt evolution — `thiserror` for errors, `DispatchOutcome` enum, safe casts, DI config injection. **V110**: Cross-ecosystem absorption — `#[expect(reason)]` migration, structured tracing, toadStool direct dispatch, dual-format capability parsing. **V109**: Zero-panic validation binaries. **V108**: AGPL-3.0-or-later.
+**Date**: March 17, 2026 | **License**: AGPL-3.0-or-later
+**Status**: V114 — 39 modules, 35 experiments, 715 Rust tests + 287 Python provenance tests, 395/395 validation checks (340 core + 55 NUCLEUS) + 140 metalForge checks, 102 active barracuda delegations (61 CPU + 41 GPU) — synced against barraCuda v0.3.5, toadStool S156+, coralReef Iteration 52+. Three-tier parity proven: 29/29 validation binaries PASS at all three tiers. **V114**: Cross-ecosystem deep absorption — `safe_cast` expansion (neuralSpring `usize_u32`/`f64_f32`), `BiomeOsError::is_recoverable()/is_retriable()` (airSpring/wetSpring), `health.liveness`/`health.readiness` probes (wetSpring/airSpring/healthSpring), `resilient_call()` (CircuitBreaker + RetryPolicy), `extract_rpc_result()` (ludoSpring/healthSpring), `FAMILY_ID`-aware discovery (toadStool/songBird), zero `eprintln!` in prod, `NEURAL_API_SOCKET_NAMES` centralization, bare `as` cast → `cast::` helpers, `.expect()` → `OrExit` in all validation binaries, primal composition guidance. **V113**: GemmF64 transpose delegation, RetryPolicy + CircuitBreaker, 4-format capability parsing. **V112**: `OrExit<T>`, provenance trio. **V111**: Deep debt — `thiserror`, `DispatchOutcome`, safe casts. **V110**: Cross-ecosystem absorption.
 
 **The gap between what models predict and what instruments measure.**
 
@@ -113,7 +113,7 @@ Clean models (other springs) → Noisy measurements (groundSpring) → Adapted m
 ### Rust Phase 1
 
 ```bash
-cargo test --workspace                         # 618 unit + 24 integration tests, all PASS
+cargo test --workspace                         # 715 tests, all PASS
 cargo test --workspace --features biomeos      # NUCLEUS client active
 cargo test --workspace --features barracuda-gpu # GPU dispatch active
 cargo clippy --workspace --all-targets -- -D warnings -W clippy::pedantic -W clippy::nursery
@@ -162,6 +162,7 @@ cargo run --bin validate-precision-drift
 cargo run --bin validate-size-convergence
 cargo run --bin validate-vendor-parity
 cargo run --bin validate-tissue-anderson
+cargo run --bin validate-et0-methods
 
 # NUCLEUS / biomeOS validation (requires biomeos feature, NUCLEUS optional)
 cargo run --features biomeos --bin validate-real-ghcnd-et0
@@ -357,7 +358,8 @@ groundSpring/
 │   ├── precision_drift/            # Exp 025: f32 vs f64 precision drift
 │   ├── size_convergence/           # Exp 026: System-size convergence
 │   ├── vendor_parity/              # Exp 027: GPU vendor parity
-│   └── npu_anderson/               # Exp 028: NPU Anderson regime classification
+│   ├── npu_anderson/               # Exp 028: NPU Anderson regime classification
+│   └── et0_methods/                # Exp 034: Multi-method ET₀ cross-validation
 ├── crates/
 │   ├── groundspring/               # Phase 1 Rust library (39 modules incl. esn, lanczos, tissue_anderson, biomeos, nestgate, npu, primal_names)
 │   └── groundspring-validate/      # 34 validation binaries (hotSpring pattern)

crates/groundspring-validate/src/bench_cpu_vs_gpu.rs

@@ -415,7 +415,13 @@ fn bench_tikhonov(iters: u32) -> BenchEntry {
 
     let cpu_ms = bench(
         || {
-            let r = groundspring::spectral_recon::tikhonov_solve(&kernel, &g, 1e-6, nt, nw);
+            let r = groundspring::spectral_recon::tikhonov_solve(
+                &kernel,
+                &g,
+                groundspring::spectral_recon::LAMBDA_NOISY,
+                nt,
+                nw,
+            );
             std::hint::black_box(r);
         },
         iters * 10,

crates/groundspring-validate/src/lib.rs

@@ -362,7 +362,6 @@ pub fn bool_field(v: &Value, key: &str) -> bool {
 #[cfg(test)]
 #[expect(
     clippy::unwrap_used,
-    clippy::expect_used,
     reason = "test assertions use unwrap/expect for clarity"
 )]
 mod tests {

crates/groundspring-validate/src/validate_aggregate_stability.rs

@@ -190,11 +190,6 @@ fn main() {
 }
 
 #[cfg(test)]
-#[expect(
-    clippy::unwrap_used,
-    clippy::expect_used,
-    reason = "test assertions use unwrap/expect for clarity"
-)]
 mod tests {
     #[test]
     fn validation_passes() {

crates/groundspring-validate/src/validate_anderson.rs

@@ -267,11 +267,6 @@ fn main() {
 }
 
 #[cfg(test)]
-#[expect(
-    clippy::unwrap_used,
-    clippy::expect_used,
-    reason = "test assertions use unwrap/expect for clarity"
-)]
 mod tests {
     #[test]
     fn validation_passes() {

crates/groundspring-validate/src/validate_band_edge.rs

@@ -15,8 +15,8 @@ use groundspring::band_structure::{
 use groundspring::transport::tridiag_eigh;
 use groundspring::validate::ValidationHarness;
 use groundspring_validate::{
-    BenchResult, f64_field, get_f64_range, get_f64_vec, parse_benchmark, print_provenance_header,
-    usize_field,
+    BenchResult, OrExit, f64_field, get_f64_range, get_f64_vec, parse_benchmark,
+    print_provenance_header, usize_field,
 };
 use serde_json::Value;
 
@@ -141,25 +141,14 @@ fn run() -> i32 {
     let exp = &bench["expected_results"];
 
     let t_hop = f64_field(model, "hopping");
-    let Ok(pot_2) = get_f64_vec(model, "period_2_potential") else {
-        eprintln!("FATAL: missing period_2_potential");
-        return 1;
-    };
-    let Ok(pot_3) = get_f64_vec(model, "period_3_potential") else {
-        eprintln!("FATAL: missing period_3_potential");
-        return 1;
-    };
+    let pot_2 = get_f64_vec(model, "period_2_potential").or_exit("missing period_2_potential");
+    let pot_3 = get_f64_vec(model, "period_3_potential").or_exit("missing period_3_potential");
     let n_scan = usize_field(model, "n_energy_scan");
-    let Ok((e_lo, e_hi)) = get_f64_range(&model["energy_range"]) else {
-        eprintln!("FATAL: missing energy_range");
-        return 1;
-    };
+    let (e_lo, e_hi) = get_f64_range(&model["energy_range"]).or_exit("missing energy_range");
     let n_periods = usize_field(model, "n_periods_finite");
 
-    let Ok(dvs) = get_f64_vec(model, "period_2_gap_widths_to_test") else {
-        eprintln!("FATAL: missing period_2_gap_widths_to_test");
-        return 1;
-    };
+    let dvs = get_f64_vec(model, "period_2_gap_widths_to_test")
+        .or_exit("missing period_2_gap_widths_to_test");
 
     let ctx = BandCtx {
         t_hop,
@@ -170,10 +159,8 @@ fn run() -> i32 {
         n_scan,
     };
 
-    if let Err(e) = validate_free_and_periodic(&mut h, &ctx, pred, exp) {
-        eprintln!("FATAL: benchmark field error: {e}");
-        return 1;
-    }
+    validate_free_and_periodic(&mut h, &ctx, pred, exp)
+        .or_exit("benchmark field error in free/periodic validation");
     validate_proportionality_and_finite(&mut h, &ctx, exp, n_periods, &dvs);
 
     println!("\n--- Part 6: Determinism ---");
@@ -192,11 +179,6 @@ fn main() {
 }
 
 #[cfg(test)]
-#[expect(
-    clippy::unwrap_used,
-    clippy::expect_used,
-    reason = "test assertions use unwrap/expect for clarity"
-)]
 mod tests {
     #[test]
     fn validation_passes() {

crates/groundspring-validate/src/validate_bistable.rs

@@ -232,11 +232,6 @@ fn main() {
 }
 
 #[cfg(test)]
-#[expect(
-    clippy::unwrap_used,
-    clippy::expect_used,
-    reason = "test assertions use unwrap/expect for clarity"
-)]
 mod tests {
     #[test]
     fn validation_passes() {