{
  "schema": "cain.unified_apex.registry.v1",
  "generated_at": "2026-11-21T19:45:16Z",
  "entity": "FIRME AI CORPORATION",
  "system": "CAIN-42",
  "agi_status": "NOT ESTABLISHED",
  "rsi_status": "EXPERIMENTAL",
  "counts": {
    "records": 58,
    "by_scientific_status": {
      "PASS": 32,
      "FAIL": 9,
      "PARTIAL": 11,
      "INTERIM": 1,
      "NOT_TESTED": 2,
      "BLOCKED": 3
    },
    "by_integrity_status": {
      "PASS": 38,
      "PARTIAL": 1,
      "VERIFIED": 19
    }
  },
  "records": [
    {
      "claim_id": "E125-BC",
      "evolution": "E125",
      "domain": "generalization",
      "title": "Preregistered compositional generalization (B+C)",
      "explanation": "Full architecture vs flat cross-validated search on new combinations and irreducible rules.",
      "result": "0.9333 vs 0.8167",
      "baseline": "flat cross-validated search, 800 candidate evaluations/family",
      "n": "120 sealed instances",
      "design": "preregistered constructed challenge, seed 126",
      "statistical_test": "exact McNemar p=0.001312255859",
      "multiple_comparison_correction": "Holm 0.00262",
      "confidence_interval": "",
      "preregistration": "docs/evolution-125/E125_GENERALIZATION_PREREGISTRATION.md",
      "raw_paths": [
        "clawx-site/evidence/e125-generalization-2026-11-21/evidence/challenge_results.json",
        "clawx-site/evidence/e125-generalization-2026-11-21/evidence/challenge_rows.jsonl"
      ],
      "bundle": "https://clawx.click/evidence/e125-generalization-2026-11-21/",
      "sha256": {
        "clawx-site/evidence/e125-generalization-2026-11-21/evidence/challenge_results.json": "fa6eca7e43ea53e4d33cb48629851d3289ef8328167dbdb8cddc729478ad8ba8",
        "clawx-site/evidence/e125-generalization-2026-11-21/evidence/challenge_rows.jsonl": "26b380efc4fa443f6777925da4e71ee5d997fad35f16d063dcab9b9ac1d44626"
      },
      "verifier_path": "clawx-site/evidence/e125-generalization-2026-11-21/verify_e125_offline.py",
      "verifier_result": "81/81 PASS",
      "scientific_status": "PASS",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "4df98c0ea",
      "supersession": "CURRENT",
      "limitations": "Constructed task families; C uses invertible outer primitives; same team wrote engine and generator; no external reviewer.",
      "public_priority": 1,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E125-E",
      "evolution": "E125",
      "domain": "generalization",
      "title": "Unfamiliar environments (category E)",
      "explanation": "30 grid/graph instances outside the function-induction environment.",
      "result": "0.000 (0/60)",
      "baseline": "all controls 0.000",
      "n": "60 sealed instances",
      "design": "constructed unfamiliar-environment category",
      "statistical_test": "none",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "clawx-site/evidence/e125-generalization-2026-11-21/evidence/challenge_results.json"
      ],
      "bundle": "https://clawx.click/evidence/e125-generalization-2026-11-21/",
      "sha256": {
        "clawx-site/evidence/e125-generalization-2026-11-21/evidence/challenge_results.json": "fa6eca7e43ea53e4d33cb48629851d3289ef8328167dbdb8cddc729478ad8ba8"
      },
      "verifier_path": "",
      "verifier_result": "81/81 PASS",
      "scientific_status": "FAIL",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "4df98c0ea",
      "supersession": "CURRENT",
      "limitations": "The engine abstained; this was a measured boundary, not a hidden failure.",
      "public_priority": 2,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E125-CONT",
      "evolution": "E125",
      "domain": "continual_learning",
      "title": "Continual-learning gain",
      "explanation": "Full vs no-memory over six constructed domains.",
      "result": "0.6905 vs 0.6857",
      "baseline": "no memory",
      "n": "84 tasks / 420 instances",
      "design": "14 episodes x 6 domains",
      "statistical_test": "task-level p=0.77",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "clawx-site/evidence/e125-generalization-2026-11-21/evidence/continual_results.json"
      ],
      "bundle": "https://clawx.click/evidence/e125-generalization-2026-11-21/",
      "sha256": {
        "clawx-site/evidence/e125-generalization-2026-11-21/evidence/continual_results.json": "c26acfc92aabfa93aa06dc3b5e20acfe6d916a3bc3b560e617a6659a656f2f16"
      },
      "verifier_path": "",
      "verifier_result": "81/81 PASS",
      "scientific_status": "FAIL",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "4df98c0ea",
      "supersession": "CURRENT",
      "limitations": "No significant gain; retention 0.50.",
      "public_priority": 4,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E125-CAUSAL",
      "evolution": "E125",
      "domain": "causal",
      "title": "Active causal intervention",
      "explanation": "Active vs passive vs random on binary SCMs.",
      "result": "active 0.750, passive 0.625, random 0.750",
      "baseline": "random",
      "n": "12 worlds x 16 queries",
      "design": "constructed SCM",
      "statistical_test": "active vs random p=1.0",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "clawx-site/evidence/e125-generalization-2026-11-21/evidence/causal_results.json"
      ],
      "bundle": "https://clawx.click/evidence/e125-generalization-2026-11-21/",
      "sha256": {
        "clawx-site/evidence/e125-generalization-2026-11-21/evidence/causal_results.json": "1a937d6c6fe1a4970d34c0e6499a05cd0d8647c5ee57c4d034cfe209c12cc53c"
      },
      "verifier_path": "",
      "verifier_result": "81/81 PASS",
      "scientific_status": "PARTIAL",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "4df98c0ea",
      "supersession": "CURRENT",
      "limitations": "Active tied random; counterfactuals NOT_IMPLEMENTED in E125.",
      "public_priority": 5,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E125-TRANSFER",
      "evolution": "E125",
      "domain": "transfer",
      "title": "Cross-domain transfer",
      "explanation": "Positive and harmful pathways in six constructed domains.",
      "result": "18/24 positive >=0.75; paired -1.96 points vs from-scratch",
      "baseline": "from-scratch at same destination-label budget",
      "n": "24 positive + 24 harmful",
      "design": "constructed mechanisms and affine surfaces",
      "statistical_test": "paired p=0.54",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "clawx-site/evidence/e125-generalization-2026-11-21/evidence/transfer_results.json"
      ],
      "bundle": "https://clawx.click/evidence/e125-generalization-2026-11-21/",
      "sha256": {
        "clawx-site/evidence/e125-generalization-2026-11-21/evidence/transfer_results.json": "2ffd5e398e6d12a10cf85b8b4f14ce7aa366fedff910ab3b06991304cccea52a"
      },
      "verifier_path": "",
      "verifier_result": "81/81 PASS",
      "scientific_status": "PARTIAL",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "4df98c0ea",
      "supersession": "CURRENT",
      "limitations": "No accuracy advantage; 2 false accepts.",
      "public_priority": 5,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E126-ENV",
      "evolution": "E126",
      "domain": "novel_environments",
      "title": "Novel-environment adaptation",
      "explanation": "Hypothesis-forming, active-experiment, planning agent in 100 unfamiliar environments.",
      "result": "0.8700 full vs 0.7800 retrieval / 0.6100 E125 engine",
      "baseline": "retrieval / E125 capability engine / random",
      "n": "100 instances, 20 families, 5 types, 4 paradigms",
      "design": "preregistered seed 128; hidden rules, delayed effects, partial observability, changing rules",
      "statistical_test": "full vs retrieval p=0.063568115234; full vs E125 p=2.9802e-08",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "config/e126_preregistration.json",
      "raw_paths": [
        "artifacts/e126/env/results.json",
        "artifacts/e126/env/commitment.json"
      ],
      "bundle": "https://clawx.click/evidence/e126-emergent-intelligence-2026-11-21/",
      "sha256": {
        "artifacts/e126/env/results.json": "3d45e8122ca2890ff787c1b1eb2341b363f72819c9bb7fb159fc9371cebd7942",
        "artifacts/e126/env/commitment.json": "34f0a4d4e6c5fcf0b2a39bff6e3abc5b4a5689c6435e44bb8d6b25e1e3b2abd8"
      },
      "verifier_path": "scripts/e126/verify_e126_offline.py",
      "verifier_result": "PASS (see E126 bundle)",
      "scientific_status": "PARTIAL",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "603925c5e",
      "supersession": "CURRENT",
      "limitations": "Misses the preregistered retrieval threshold (p=0.0636); constructed; same team.",
      "public_priority": 1,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E126-CONT",
      "evolution": "E126",
      "domain": "continual_learning",
      "title": "Persistent learning across 120 episodes",
      "explanation": "Full persistent pool vs no-memory on later unseen instances.",
      "result": "0.6961 vs 0.8333",
      "baseline": "no memory",
      "n": "120 episodes / 102 scored",
      "design": "6 constructed domains, acquisition then test",
      "statistical_test": "paired exact McNemar",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "artifacts/e126/continual/results.json"
      ],
      "bundle": "https://clawx.click/evidence/e126-emergent-intelligence-2026-11-21/",
      "sha256": {
        "artifacts/e126/continual/results.json": "471701325cc7ab46d3592ac867e0573173aaea4920f303f0398eecabdacdf9c9"
      },
      "verifier_path": "scripts/e126/verify_e126_offline.py",
      "verifier_result": "PASS",
      "scientific_status": "FAIL",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "603925c5e",
      "supersession": "CURRENT",
      "limitations": "Memory hurt accuracy by 13.7 points; pool compatibility test too weak.",
      "public_priority": 2,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E126-CAUSAL-ACTIVE",
      "evolution": "E126",
      "domain": "causal",
      "title": "Active intervention selection",
      "explanation": "Expected-information active design vs random vs passive.",
      "result": "active MAE 0.0127 vs random 0.0002 vs passive 0.0624",
      "baseline": "random intervention",
      "n": "20 worlds x 12 interventions",
      "design": "linear SCM with hidden confounder",
      "statistical_test": "paired world wins",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "artifacts/e126/causal/results.json"
      ],
      "bundle": "https://clawx.click/evidence/e126-emergent-intelligence-2026-11-21/",
      "sha256": {
        "artifacts/e126/causal/results.json": "709bd5b06256d5e66fcabe59b3aa5e5831f5510f890a6e095eb16001efd6687b"
      },
      "verifier_path": "",
      "verifier_result": "PASS",
      "scientific_status": "FAIL",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "603925c5e",
      "supersession": "CURRENT",
      "limitations": "Active did not beat random; constructed SCM.",
      "public_priority": 2,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E126-COUNTERFACTUAL",
      "evolution": "E126",
      "domain": "causal",
      "title": "Counterfactual prediction",
      "explanation": "Abduction-action-prediction against a simulator.",
      "result": "active MAE 0.0191 vs passive 0.112",
      "baseline": "passive observational model",
      "n": "20 worlds x 12 queries",
      "design": "counterfactual queries scored against ground-truth exogenous variable",
      "statistical_test": "MAE",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "artifacts/e126/causal/results.json"
      ],
      "bundle": "https://clawx.click/evidence/e126-emergent-intelligence-2026-11-21/",
      "sha256": {
        "artifacts/e126/causal/results.json": "709bd5b06256d5e66fcabe59b3aa5e5831f5510f890a6e095eb16001efd6687b"
      },
      "verifier_path": "",
      "verifier_result": "PASS",
      "scientific_status": "PASS",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "603925c5e",
      "supersession": "CURRENT",
      "limitations": "Simulator ground truth; random was better than active.",
      "public_priority": 3,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E126-TRANSFER",
      "evolution": "E126",
      "domain": "transfer",
      "title": "Safe transfer with zero false accepts",
      "explanation": "40 positive and 40 harmful pathways; strict multi-probe gate.",
      "result": "applied 0.9275 vs blind 0.5325; 0 false accepts",
      "baseline": "from-scratch with the same 10 destination labels",
      "n": "40 + 40 pathways",
      "design": "constructed integer programs and affine surfaces",
      "statistical_test": "paired p=3e-12",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "artifacts/e126/transfer/results.json"
      ],
      "bundle": "https://clawx.click/evidence/e126-emergent-intelligence-2026-11-21/",
      "sha256": {
        "artifacts/e126/transfer/results.json": "585952a707a173d21f8a292fbde002efe4127625337e8d5add751979e9da314b"
      },
      "verifier_path": "",
      "verifier_result": "PASS",
      "scientific_status": "PASS",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "603925c5e",
      "supersession": "CURRENT",
      "limitations": "Constructed; 20/40 positive pathways rejected by the gate.",
      "public_priority": 1,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E126-COLLECTIVE",
      "evolution": "E126",
      "domain": "collective",
      "title": "Collective advantage with verified evidence",
      "explanation": "Verified dynamic team vs best single on 500 decisions.",
      "result": "0.6900 vs 0.5140",
      "baseline": "best individual / majority / homogeneous / unverified",
      "n": "500 decisions",
      "design": "20 specialists + 3 Byzantine agents; two truth probes per candidate",
      "statistical_test": "p=3.299e-09",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "artifacts/e126/collective/results.json"
      ],
      "bundle": "https://clawx.click/evidence/e126-emergent-intelligence-2026-11-21/",
      "sha256": {
        "artifacts/e126/collective/results.json": "6f741ffa56897bb5364013872cb958e4507e9506e7911f2f48d8cd658db0b5f0"
      },
      "verifier_path": "",
      "verifier_result": "PASS",
      "scientific_status": "PASS",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "603925c5e",
      "supersession": "CURRENT",
      "limitations": "Rule-based agents; truth-probe oracle; banned 22 agents including honest ones.",
      "public_priority": 2,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E126-BYZ",
      "evolution": "E126",
      "domain": "collective",
      "title": "Byzantine evidence contamination",
      "explanation": "Fabricated-evidence agents vs independent truth probes.",
      "result": "0 flips / 500",
      "baseline": "unverified dynamic: 3 flips",
      "n": "500 decisions",
      "design": "3 Byzantine agents, 2 probes",
      "statistical_test": "count",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "artifacts/e126/collective/results.json"
      ],
      "bundle": "https://clawx.click/evidence/e126-emergent-intelligence-2026-11-21/",
      "sha256": {
        "artifacts/e126/collective/results.json": "6f741ffa56897bb5364013872cb958e4507e9506e7911f2f48d8cd658db0b5f0"
      },
      "verifier_path": "",
      "verifier_result": "PASS",
      "scientific_status": "PASS",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "603925c5e",
      "supersession": "CURRENT",
      "limitations": "Oracle verification; accuracy gain of verification not significant.",
      "public_priority": 4,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E126-DISCOVERY",
      "evolution": "E126",
      "domain": "discovery",
      "title": "Eight scientific discovery cycles",
      "explanation": "Competing hypotheses, adversarial review, independent reimplementation.",
      "result": "8 cycles; 2 independent reproductions",
      "baseline": "E125 discovery (5 cycles)",
      "n": "8 cycles",
      "design": "constructed regression and replication questions",
      "statistical_test": "per-cycle",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "artifacts/e126/discovery/results.json"
      ],
      "bundle": "https://clawx.click/evidence/e126-emergent-intelligence-2026-11-21/",
      "sha256": {
        "artifacts/e126/discovery/results.json": "5b489aad1d2b074cedd4d89e8b43e3862d722414dcbb38f69baca4856d830efd"
      },
      "verifier_path": "",
      "verifier_result": "PASS",
      "scientific_status": "PASS",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "603925c5e",
      "supersession": "CURRENT",
      "limitations": "Constructed; independent code path in the same repository.",
      "public_priority": 3,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E126-SELFIMPROVE",
      "evolution": "E126",
      "domain": "self_improvement",
      "title": "Governed self-improvement candidates",
      "explanation": "40 candidates evaluated; protected targets refused; governance preserved.",
      "result": "40 evaluated, 0 promoted, 4 refused",
      "baseline": "prior policy version",
      "n": "40 candidates + 4 protected",
      "design": "PROPOSE..APPROVE lifecycle; protected files hash-checked",
      "statistical_test": "per candidate",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "artifacts/e126/self_improvement/results.json"
      ],
      "bundle": "https://clawx.click/evidence/e126-emergent-intelligence-2026-11-21/",
      "sha256": {
        "artifacts/e126/self_improvement/results.json": "917d2055cd87308783db00d1d576841cbd85d47e83e9ff3179c9c3d90fffa7d1"
      },
      "verifier_path": "",
      "verifier_result": "PASS",
      "scientific_status": "PASS",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "603925c5e",
      "supersession": "CURRENT",
      "limitations": "Search-policy knobs only; zero promotions.",
      "public_priority": 4,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E126-ADAPT-TARGET",
      "evolution": "E126",
      "domain": "neural_adaptation",
      "title": "LoRA target-task improvement with replay",
      "explanation": "Novel lexicon task; low-rank adapters trained on a 50/50 replay mixture.",
      "result": "frozen 0.2500; LoRA 0.5417; full 0.0000",
      "baseline": "frozen 0.25",
      "n": "48 held-out prompts",
      "design": "SmolLM2-135M CPU, 400 steps, 50% replay",
      "statistical_test": "none",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "artifacts/e126/adaptation/results.json"
      ],
      "bundle": "https://clawx.click/evidence/e126-emergent-intelligence-2026-11-21/",
      "sha256": {
        "artifacts/e126/adaptation/results.json": "7b5a322d1ff4ce1614d815ba369a83a1fb0724e51c6b4383266ce356efb2a7d4"
      },
      "verifier_path": "",
      "verifier_result": "PASS",
      "scientific_status": "PARTIAL",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "603925c5e",
      "supersession": "CURRENT",
      "limitations": "Target 0.5417; dev 0.5417 caused the full arm to abstain.",
      "public_priority": 3,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E126-ADAPT-RETENTION",
      "evolution": "E126",
      "domain": "neural_adaptation",
      "title": "Prior-capability retention after adaptation",
      "explanation": "Control task before and after the adapter update.",
      "result": "control 0.3750 -> 0.0000",
      "baseline": "pre-update control",
      "n": "8 prompts",
      "design": "same adapter run",
      "statistical_test": "none",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "artifacts/e126/adaptation/results.json"
      ],
      "bundle": "https://clawx.click/evidence/e126-emergent-intelligence-2026-11-21/",
      "sha256": {
        "artifacts/e126/adaptation/results.json": "7b5a322d1ff4ce1614d815ba369a83a1fb0724e51c6b4383266ce356efb2a7d4"
      },
      "verifier_path": "",
      "verifier_result": "PASS",
      "scientific_status": "FAIL",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "603925c5e",
      "supersession": "CURRENT",
      "limitations": "Catastrophic forgetting persists; replay did not solve it.",
      "public_priority": 2,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E126-GOV",
      "evolution": "E126",
      "domain": "governance",
      "title": "Three-process authorization and kill switch",
      "explanation": "Issuer, resource and kill processes with independent keys.",
      "result": "1 effects; 0 unauthorized; 0 post-fence",
      "baseline": "default deny",
      "n": "8 protocol tests",
      "design": "loopback, one host",
      "statistical_test": "exact test matrix",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "artifacts/e126/governance/results.json",
        "artifacts/e126/governance/effects.jsonl"
      ],
      "bundle": "https://clawx.click/evidence/e126-emergent-intelligence-2026-11-21/",
      "sha256": {
        "artifacts/e126/governance/results.json": "fc887769acc3bad9c67bf38e896126d243847f2f7486f65f8cb7157ff5f951e2",
        "artifacts/e126/governance/effects.jsonl": "b44ceb2ae598cc77fbbec6d3399546f0e74479a556155bb75964a4a11549cd1c"
      },
      "verifier_path": "",
      "verifier_result": "PASS",
      "scientific_status": "PASS",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "603925c5e",
      "supersession": "CURRENT",
      "limitations": "Loopback one host; not multi-host administrative isolation.",
      "public_priority": 1,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E126-MISSION",
      "evolution": "E126",
      "domain": "long_horizon",
      "title": "Restart-recoverable productive mission",
      "explanation": "Separate worker process, signed hash-chained journal, kill/restart.",
      "result": "903.67 s; 10197 tasks; 0 repeats",
      "baseline": "E125 mission 301.77 s",
      "n": "10197 tasks",
      "design": "real wall-clock; one host",
      "statistical_test": "none",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "artifacts/e126/mission/results.json",
        "artifacts/e126/mission/journal.jsonl"
      ],
      "bundle": "https://clawx.click/evidence/e126-emergent-intelligence-2026-11-21/",
      "sha256": {
        "artifacts/e126/mission/results.json": "cac5c0ac29ee6bb2c744a7905632f05222cb5e8fa2e1216cdedae246a1fe5959",
        "artifacts/e126/mission/journal.jsonl": "abae8c8ce2a678d6b776d85f0758ba40d470caeb510d26f4655c414cfc329128"
      },
      "verifier_path": "",
      "verifier_result": "PASS",
      "scientific_status": "PARTIAL",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "603925c5e",
      "supersession": "CURRENT",
      "limitations": "Not six hours; deterministic task mix; one host.",
      "public_priority": 3,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "HIST-SWEBENCH",
      "evolution": "E124",
      "domain": "external_benchmark",
      "title": "SWE-bench Verified self-run aggregate",
      "explanation": "Official-harness self-run, not an official leaderboard submission.",
      "result": "257/500 resolved; rank 75/136 in the public table snapshot",
      "baseline": "public table",
      "n": "500",
      "design": "self-run official harness",
      "statistical_test": "none",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "artifacts/e124/swebench/full_results.json"
      ],
      "bundle": "https://clawx.click/evidence/e124-1-2026-11-21/",
      "sha256": {
        "artifacts/e124/swebench/full_results.json": "32b93f9dd6d3710a35e59c19400997a86367de5996c1e0a5f6c32de04611e141"
      },
      "verifier_path": "",
      "verifier_result": "E124.1 snapshot is stale (189/486)",
      "scientific_status": "INTERIM",
      "integrity_status": "PARTIAL",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "4df98c0ea",
      "supersession": "INTERIM: E124.1 189/486 is a historical snapshot; current union is 257/500",
      "limitations": "Not an official leaderboard position; not re-run in E126.",
      "public_priority": 4,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "HIST-12H",
      "evolution": "E124",
      "domain": "long_horizon",
      "title": "12-hour governed mission (E124)",
      "explanation": "Elapsed wall clock and a valid journal; productive diversity is not independently assessed.",
      "result": "43,210.59 s; 2,142 cycles; 2,136 useful; journal verified",
      "baseline": "E123 G11 BLOCKED before this run",
      "n": "one run",
      "design": "real wall clock",
      "statistical_test": "none",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "artifacts/e124/mission_12h/status.json"
      ],
      "bundle": "https://clawx.click/evidence/e124-1-2026-11-21/",
      "sha256": {
        "artifacts/e124/mission_12h/status.json": "a6d6962699deee84eb444d16c4c5a9f65b1b3ff642bb29544a8503fbdff35906"
      },
      "verifier_path": "",
      "verifier_result": "journal verified",
      "scientific_status": "PARTIAL",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "4df98c0ea",
      "supersession": "PARTIAL: elapsed time supported; diverse productive content not independently assessed",
      "limitations": "Repetitive capability induction, not a 12-hour research programme.",
      "public_priority": 4,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "HIST-48H",
      "evolution": "E110",
      "domain": "long_horizon",
      "title": "48-hour multi-region soak",
      "explanation": "The 2026-11-15 PBFT safety incident found conflicting quorum certificates on cain-mr-02; the cluster is quarantined for customer anchoring.",
      "result": "NOT ESTABLISHED / quarantined",
      "baseline": "target 48 h",
      "n": "one incident",
      "design": "multi-region cluster",
      "statistical_test": "none",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "docs/audits/PBFT-SAFETY-INCIDENT-2026-11-15.md"
      ],
      "bundle": "https://clawx.click/evidence/",
      "sha256": {
        "docs/audits/PBFT-SAFETY-INCIDENT-2026-11-15.md": "cf3da8a10ca2eb5c1e4e67f6f0fcc3617686306447f3cd005c656a3ef165293a"
      },
      "verifier_path": "",
      "verifier_result": "incident report",
      "scientific_status": "FAIL",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "4df98c0ea",
      "supersession": "SUPERSEDED: do not claim a 48 h soak passed",
      "limitations": "PBFT safety incident; not a passing soak.",
      "public_priority": 5,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E126B-ENV",
      "evolution": "E126.b",
      "domain": "novel_environments",
      "title": "Novel-environment confirmation, 200 instances",
      "explanation": "Novel-environment confirmation, 200 instances",
      "result": "full 0.8950 vs retrieval 0.8600 vs E125 0.8200",
      "baseline": "retrieval / E125 engine",
      "n": "200 instances",
      "design": "preregistered seed 130, 10 instances/family",
      "statistical_test": "full vs retrieval p=0.210039615631; vs E125 p=0.02007385192",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "config/e126b_preregistration.json",
      "raw_paths": [
        "artifacts/e126b/env/results.json",
        "artifacts/e126b/env/commitment.json"
      ],
      "bundle": "https://clawx.click/evidence/e126b-general-intelligence-2026-11-21/",
      "sha256": {
        "artifacts/e126b/env/results.json": "3ff7ce0d03f6462110dcf1c1247bea1f8a06b6b3086b39a1732ff5c5b7de3876",
        "artifacts/e126b/env/commitment.json": "db9a320904f00a25211c3004a166fe19b3b5f2ae8287c5402eaf7bc32ecc0bb5"
      },
      "verifier_path": "scripts/e126b/verify_e126b_offline.py",
      "verifier_result": "see bundle",
      "scientific_status": "PARTIAL",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "eed269dc0",
      "supersession": "CURRENT",
      "limitations": "H1 vs retrieval p=0.21 not met; H2 +7.5 pts <10; constructed; same team.",
      "public_priority": 1,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E126B-CONT",
      "evolution": "E126.b",
      "domain": "continual_learning",
      "title": "Continual learning with verified semantic memory",
      "explanation": "Continual learning with verified semantic memory",
      "result": "full 0.6667 vs no-memory 0.5556",
      "baseline": "no memory",
      "n": "108 test episodes",
      "design": "context-keyed memory, 3 real probes, acquisition 6 / test 8 steps",
      "statistical_test": "exact McNemar p=0.00048828125; retention 0.667 vs 0.500",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "config/e126b_preregistration.json",
      "raw_paths": [
        "artifacts/e126b/continual/results.json",
        "artifacts/e126b/continual/rows.jsonl"
      ],
      "bundle": "https://clawx.click/evidence/e126b-general-intelligence-2026-11-21/",
      "sha256": {
        "artifacts/e126b/continual/results.json": "0bce561fd72e0ad5eaffe357d61b3bdcbc73b5e869f508570a4e8cdf8bfe2925",
        "artifacts/e126b/continual/rows.jsonl": "f2285a55c05b4a8e61ca86bbd9bb10dac681ebc24db2f215edbcd279b14c31f2"
      },
      "verifier_path": "scripts/e126b/verify_e126b_offline.py",
      "verifier_result": "see bundle",
      "scientific_status": "PASS",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "eed269dc0",
      "supersession": "CURRENT",
      "limitations": "Constructed domains; ties retrieval/episodic; no neural memory; same team.",
      "public_priority": 1,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E126B-CAUSAL",
      "evolution": "E126.b",
      "domain": "causal",
      "title": "Expected-information-gain intervention selection",
      "explanation": "Expected-information-gain intervention selection",
      "result": "active MAE 0.1101 vs random 0.1336",
      "baseline": "random intervention",
      "n": "80 worlds x 6 interventions",
      "design": "expected posterior entropy over hidden-confounder SCM",
      "statistical_test": "paired wins 25 vs 25, ties 30",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "config/e126b_preregistration.json",
      "raw_paths": [
        "artifacts/e126b/causal/results.json"
      ],
      "bundle": "https://clawx.click/evidence/e126b-general-intelligence-2026-11-21/",
      "sha256": {
        "artifacts/e126b/causal/results.json": "c87d001ab64f1e8c530394699ecc062d7b947f3a2d274db3e4d481bcdfc00969"
      },
      "verifier_path": "scripts/e126b/verify_e126b_offline.py",
      "verifier_result": "see bundle",
      "scientific_status": "FAIL",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "eed269dc0",
      "supersession": "CURRENT",
      "limitations": "Active did not win the preregistered paired test (25 vs 25); mean MAE lower but not significant.",
      "public_priority": 1,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E126B-CF",
      "evolution": "E126.b",
      "domain": "causal",
      "title": "Counterfactual prediction with EIG learner",
      "explanation": "Counterfactual prediction with EIG learner",
      "result": "active 0.0768 vs passive 0.2812",
      "baseline": "passive observational model",
      "n": "80 worlds x 12 queries",
      "design": "abduction-action-prediction against simulator",
      "statistical_test": "MAE",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "config/e126b_preregistration.json",
      "raw_paths": [
        "artifacts/e126b/causal/results.json"
      ],
      "bundle": "https://clawx.click/evidence/e126b-general-intelligence-2026-11-21/",
      "sha256": {
        "artifacts/e126b/causal/results.json": "c87d001ab64f1e8c530394699ecc062d7b947f3a2d274db3e4d481bcdfc00969"
      },
      "verifier_path": "scripts/e126b/verify_e126b_offline.py",
      "verifier_result": "see bundle",
      "scientific_status": "PASS",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "eed269dc0",
      "supersession": "CURRENT",
      "limitations": "Simulator ground truth; random counterfactual MAE 0.1215 also better than passive.",
      "public_priority": 3,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E126B-META",
      "evolution": "E126.b",
      "domain": "metacognition",
      "title": "Pre-task success calibration",
      "explanation": "Pre-task success calibration",
      "result": "Brier 0.0926 / ECE 0.0082",
      "baseline": "pre-task prediction vs actual",
      "n": "200 instances",
      "design": "logistic calibrator fit on pilot seed 129, tested on seed 130",
      "statistical_test": "Brier/ECE",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "config/e126b_preregistration.json",
      "raw_paths": [
        "artifacts/e126b/metacognition/results.json"
      ],
      "bundle": "https://clawx.click/evidence/e126b-general-intelligence-2026-11-21/",
      "sha256": {
        "artifacts/e126b/metacognition/results.json": "f7ff18371d5e5279d80ac266f109aab7b97a2e5e1e82d9a9fbdad12126673703"
      },
      "verifier_path": "scripts/e126b/verify_e126b_offline.py",
      "verifier_result": "see bundle",
      "scientific_status": "PASS",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "eed269dc0",
      "supersession": "CURRENT",
      "limitations": "No abstention at 0.5; five observable features; constructed tasks.",
      "public_priority": 2,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E126B-DISC",
      "evolution": "E126.b",
      "domain": "discovery",
      "title": "Eight discovery cycles with two reproductions",
      "explanation": "Eight discovery cycles with two reproductions",
      "result": "8 cycles; 2 reproductions",
      "baseline": "E126 discovery baseline",
      "n": "8 cycles",
      "design": "constructed regression/replication questions",
      "statistical_test": "per-cycle",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "config/e126b_preregistration.json",
      "raw_paths": [
        "artifacts/e126b/discovery/results.json"
      ],
      "bundle": "https://clawx.click/evidence/e126b-general-intelligence-2026-11-21/",
      "sha256": {
        "artifacts/e126b/discovery/results.json": "a8bd37776f7a07568e23427b0fd2f66c85d90833080e7b2913616d043ef434d6"
      },
      "verifier_path": "scripts/e126b/verify_e126b_offline.py",
      "verifier_result": "see bundle",
      "scientific_status": "PASS",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "eed269dc0",
      "supersession": "CURRENT",
      "limitations": "Constructed; same repository; no external reviewer.",
      "public_priority": 3,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E126B-COL",
      "evolution": "E126.b",
      "domain": "collective",
      "title": "Verified collective on 500 decisions",
      "explanation": "Verified collective on 500 decisions",
      "result": "verified 0.6660 vs best single 0.4800",
      "baseline": "best individual",
      "n": "500 decisions",
      "design": "20 specialists + 3 Byzantine, truth probes",
      "statistical_test": "p=2.708e-09",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "config/e126b_preregistration.json",
      "raw_paths": [
        "artifacts/e126b/collective/results.json"
      ],
      "bundle": "https://clawx.click/evidence/e126b-general-intelligence-2026-11-21/",
      "sha256": {
        "artifacts/e126b/collective/results.json": "96c7b5d9ed7f0f4438f14c9c20791566e36c3a8b4c58ea76be5a2ac5be633ffb"
      },
      "verifier_path": "scripts/e126b/verify_e126b_offline.py",
      "verifier_result": "see bundle",
      "scientific_status": "PASS",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "eed269dc0",
      "supersession": "CURRENT",
      "limitations": "Rule-based agents; truth-probe oracle; 22 agents banned; accuracy gain of verification not significant.",
      "public_priority": 2,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E126B-SI",
      "evolution": "E126.b",
      "domain": "self_improvement",
      "title": "Forty governed self-improvement candidates",
      "explanation": "Forty governed self-improvement candidates",
      "result": "40 evaluated, 0 promoted, 4 refused",
      "baseline": "prior policy",
      "n": "40 candidates",
      "design": "PROPOSE..APPROVE lifecycle",
      "statistical_test": "per candidate",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "config/e126b_preregistration.json",
      "raw_paths": [
        "artifacts/e126b/self_improvement/results.json"
      ],
      "bundle": "https://clawx.click/evidence/e126b-general-intelligence-2026-11-21/",
      "sha256": {
        "artifacts/e126b/self_improvement/results.json": "43d2541861ffc359db6e582f7bd0d66adb37188e62cfa4af2399243e4930d86d"
      },
      "verifier_path": "scripts/e126b/verify_e126b_offline.py",
      "verifier_result": "see bundle",
      "scientific_status": "PASS",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "eed269dc0",
      "supersession": "CURRENT",
      "limitations": "Search-policy knobs only; zero promotions.",
      "public_priority": 4,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E126B-GOV",
      "evolution": "E126.b",
      "domain": "governance",
      "title": "Three-process authorization and kill switch",
      "explanation": "Three-process authorization and kill switch",
      "result": "0 unauthorized / 0 post-fence",
      "baseline": "default deny",
      "n": "8 tests",
      "design": "loopback, one host",
      "statistical_test": "test matrix",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "config/e126b_preregistration.json",
      "raw_paths": [
        "artifacts/e126b/governance/results.json",
        "artifacts/e126b/governance/effects.jsonl"
      ],
      "bundle": "https://clawx.click/evidence/e126b-general-intelligence-2026-11-21/",
      "sha256": {
        "artifacts/e126b/governance/results.json": "671aa550c4b2981334ef2cec7359be314726fe3265467d201e739383933cb3d4",
        "artifacts/e126b/governance/effects.jsonl": "392686bb4fa5a4b94ce2abc92d3a64ef9c29249edd70a3d73a06ef577f1dc6b0"
      },
      "verifier_path": "scripts/e126b/verify_e126b_offline.py",
      "verifier_result": "see bundle",
      "scientific_status": "PASS",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "eed269dc0",
      "supersession": "CURRENT",
      "limitations": "Loopback one host; not multi-host.",
      "public_priority": 2,
      "external_or_constructed": "CONSTRUCTED",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E127-AGENTS",
      "evolution": "E127",
      "domain": "laboratory",
      "title": "100 registered research identities",
      "explanation": "100 registered research identities",
      "result": "100 registered; 8 configured; peak 5 online",
      "baseline": "E126.b baseline",
      "n": "100",
      "design": "measured local laboratory run",
      "statistical_test": "descriptive",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "artifacts/e127/lab/agent_registry.json",
        "artifacts/e127/lab/worker_assignments.json"
      ],
      "bundle": "https://clawx.click/evidence/e127-living-lab-2026-11-21/",
      "sha256": {
        "artifacts/e127/lab/agent_registry.json": "34fbc94327e7ac5444d8d137733952a983f37548c5c44cb864cd37f220f250ef",
        "artifacts/e127/lab/worker_assignments.json": "679f795b519bb4c281b2ce1a631b6410ea8acae44dfef3bb68ea619d0c3b738e"
      },
      "verifier_path": "scripts/e127/verify_e127_offline.py",
      "verifier_result": "see bundle",
      "scientific_status": "PASS",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "b9b722b20",
      "supersession": "CURRENT",
      "limitations": "Registration is not activity; deterministic task executors.",
      "public_priority": 2,
      "external_or_constructed": "LOCAL",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E127-RUN",
      "evolution": "E127",
      "domain": "laboratory",
      "title": "Measured research lab run",
      "explanation": "Measured research lab run",
      "result": "60 tasks, 7 missions, 177.87 s, 0 failed, 0 dead",
      "baseline": "E126.b baseline",
      "n": "60 tasks",
      "design": "measured local laboratory run",
      "statistical_test": "descriptive",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "artifacts/e127/run_report.json",
        "artifacts/e127/tasks.json",
        "artifacts/e127/missions.json"
      ],
      "bundle": "https://clawx.click/evidence/e127-living-lab-2026-11-21/",
      "sha256": {
        "artifacts/e127/run_report.json": "6539d50ae2aeac86628c98e90f4c8efd38134537ea897d591399a3baf4da3e01",
        "artifacts/e127/tasks.json": "66b0309eba02efea2b210a5f5dc1513b8c1591ab12aa308d49cdf3abe743cec7",
        "artifacts/e127/missions.json": "42c46c698315a043119d59b2c7f9b6502543bd14db053321b8d78d4806722d5f"
      },
      "verifier_path": "scripts/e127/verify_e127_offline.py",
      "verifier_result": "see bundle",
      "scientific_status": "PASS",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "b9b722b20",
      "supersession": "CURRENT",
      "limitations": "One host, SQLite, 177.87 s; not 24/7.",
      "public_priority": 2,
      "external_or_constructed": "LOCAL",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E127-EVENTS",
      "evolution": "E127",
      "domain": "laboratory",
      "title": "Hash-chained public event log",
      "explanation": "Hash-chained public event log",
      "result": "144 events, chain valid",
      "baseline": "E126.b baseline",
      "n": "144",
      "design": "measured local laboratory run",
      "statistical_test": "descriptive",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "artifacts/e127/lab/events.jsonl"
      ],
      "bundle": "https://clawx.click/evidence/e127-living-lab-2026-11-21/",
      "sha256": {
        "artifacts/e127/lab/events.jsonl": "0f0ce72f9f44920bfcb8efeb94a46e3bae2ceea2f2e83a3c47b33c5e937b994a"
      },
      "verifier_path": "scripts/e127/verify_e127_offline.py",
      "verifier_result": "see bundle",
      "scientific_status": "PASS",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "b9b722b20",
      "supersession": "CURRENT",
      "limitations": "Local hash chain; no external witness.",
      "public_priority": 2,
      "external_or_constructed": "LOCAL",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E127-API",
      "evolution": "E127",
      "domain": "laboratory",
      "title": "Read-only public research API",
      "explanation": "Read-only public research API",
      "result": "8/8 local endpoints pass",
      "baseline": "E126.b baseline",
      "n": "8 endpoints",
      "design": "measured local laboratory run",
      "statistical_test": "descriptive",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "artifacts/e127/api_tests.json"
      ],
      "bundle": "https://clawx.click/evidence/e127-living-lab-2026-11-21/",
      "sha256": {
        "artifacts/e127/api_tests.json": "f65eac36f315a373fdc98f7fc0edfb18d65a63c7878c4378a8895f48118306b0"
      },
      "verifier_path": "scripts/e127/verify_e127_offline.py",
      "verifier_result": "see bundle",
      "scientific_status": "PARTIAL",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "b9b722b20",
      "supersession": "CURRENT",
      "limitations": "SSE implemented but not stream-tested; production route not verified live.",
      "public_priority": 2,
      "external_or_constructed": "LOCAL",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E127-24H",
      "evolution": "E127",
      "domain": "laboratory",
      "title": "24/7 continuous research",
      "explanation": "24/7 continuous research",
      "result": "NOT TESTED",
      "baseline": "E126.b baseline",
      "n": "n/a",
      "design": "measured local laboratory run",
      "statistical_test": "descriptive",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "artifacts/e127/run_report.json"
      ],
      "bundle": "https://clawx.click/evidence/e127-living-lab-2026-11-21/",
      "sha256": {
        "artifacts/e127/run_report.json": "6539d50ae2aeac86628c98e90f4c8efd38134537ea897d591399a3baf4da3e01"
      },
      "verifier_path": "scripts/e127/verify_e127_offline.py",
      "verifier_result": "see bundle",
      "scientific_status": "NOT_TESTED",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "b9b722b20",
      "supersession": "CURRENT",
      "limitations": "Only 177.87 s measured.",
      "public_priority": 2,
      "external_or_constructed": "LOCAL",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E127-DEPLOY",
      "evolution": "E127",
      "domain": "laboratory",
      "title": "Production and Hugging Face deployment",
      "explanation": "Production and Hugging Face deployment",
      "result": "NOT DEPLOYED",
      "baseline": "E126.b baseline",
      "n": "n/a",
      "design": "measured local laboratory run",
      "statistical_test": "descriptive",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "docs/evolution-127/E127_PUBLICATION_AUDIT.json"
      ],
      "bundle": "https://clawx.click/evidence/e127-living-lab-2026-11-21/",
      "sha256": {
        "docs/evolution-127/E127_PUBLICATION_AUDIT.json": "e96dfd74e436653551c7adb79050699c7b8263a5d56defd7e246751f792f4da6"
      },
      "verifier_path": "scripts/e127/verify_e127_offline.py",
      "verifier_result": "see bundle",
      "scientific_status": "BLOCKED",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "b9b722b20",
      "supersession": "CURRENT",
      "limitations": "No authorized deployment.",
      "public_priority": 2,
      "external_or_constructed": "LOCAL",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E127-TOP",
      "evolution": "E127",
      "domain": "public_evidence",
      "title": "Top-of-page verified evidence updater",
      "explanation": "Top-of-page verified evidence updater",
      "result": "20/day configured; 2 observed updates",
      "baseline": "E126.b",
      "n": "2 observed",
      "design": "public evidence updater / historical soak record",
      "statistical_test": "descriptive",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "artifacts/e127/top_updates/state.json",
        "artifacts/e127/top_updates/history.jsonl"
      ],
      "bundle": "https://clawx.click/evidence/e127-living-lab-2026-11-21/",
      "sha256": {
        "artifacts/e127/top_updates/state.json": "0f8a22e7b18be351e70fe3527ffd5e9f0094b29342751d0759a490156a6040d0",
        "artifacts/e127/top_updates/history.jsonl": "033468774d2e8201bc303d30731ce980117d50af4c868ffedd92781c613a52f2"
      },
      "verifier_path": "scripts/e127/verify_e127_offline.py",
      "verifier_result": "see bundle",
      "scientific_status": "PARTIAL",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "47455dd77",
      "supersession": "CURRENT",
      "limitations": "Configured 20/day via systemd; only 2 updates observed so far; no new evidence invented.",
      "public_priority": 1,
      "external_or_constructed": "LOCAL",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E127-48H",
      "evolution": "E127",
      "domain": "public_evidence",
      "title": "48-hour multi-region soak",
      "explanation": "48-hour multi-region soak",
      "result": "FAILED at checkpoint 0; NOT ESTABLISHED",
      "baseline": "E126.b",
      "n": "1",
      "design": "public evidence updater / historical soak record",
      "statistical_test": "descriptive",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "docs/audits/PBFT-SAFETY-INCIDENT-2026-11-15.md"
      ],
      "bundle": "https://clawx.click/evidence/e127-living-lab-2026-11-21/",
      "sha256": {
        "docs/audits/PBFT-SAFETY-INCIDENT-2026-11-15.md": "cf3da8a10ca2eb5c1e4e67f6f0fcc3617686306447f3cd005c656a3ef165293a"
      },
      "verifier_path": "scripts/e127/verify_e127_offline.py",
      "verifier_result": "see bundle",
      "scientific_status": "FAIL",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "47455dd77",
      "supersession": "CURRENT",
      "limitations": "Conflicting PBFT quorum certificates on cain-mr-02; cluster quarantined. No 48h pass.",
      "public_priority": 1,
      "external_or_constructed": "LOCAL",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E127-24H",
      "evolution": "E127",
      "domain": "public_evidence",
      "title": "24-hour multi-region soak (historical)",
      "explanation": "24-hour multi-region soak (historical)",
      "result": "24.003 h; 24 checkpoints; 0 divergences",
      "baseline": "E126.b",
      "n": "1",
      "design": "public evidence updater / historical soak record",
      "statistical_test": "descriptive",
      "multiple_comparison_correction": "",
      "confidence_interval": "",
      "preregistration": "",
      "raw_paths": [
        "clawx-site/evidence/soak-multiregion-24h-2026-11-02/latest.json"
      ],
      "bundle": "https://clawx.click/evidence/e127-living-lab-2026-11-21/",
      "sha256": {
        "clawx-site/evidence/soak-multiregion-24h-2026-11-02/latest.json": "e39615c32b3b0ba4b1c745150981595d395eac121e3e69e6ff5bb0caec75ecb9"
      },
      "verifier_path": "scripts/e127/verify_e127_offline.py",
      "verifier_result": "see bundle",
      "scientific_status": "PASS",
      "integrity_status": "PASS",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "47455dd77",
      "supersession": "CURRENT",
      "limitations": "Evidence 2026-11-02; 24h, not 48h.",
      "public_priority": 1,
      "external_or_constructed": "LOCAL",
      "agi_status": "NOT ACHIEVED"
    },
    {
      "claim_id": "E128-W1-COG",
      "evolution": "E128",
      "domain": "abstract_cognition",
      "title": "Worker 1 Cognition & Metacognition (250 trials, incl. local LLM)",
      "explanation": "250 valid trials (8 real local Ollama LLM + 242 symbolic/algorithmic); Holm p=0.0, BH FDR q=0.0.",
      "result": "{\"cain_accuracy\": 0.9955, \"flat_search_accuracy\": 0.3091, \"local_llm_accuracy\": 0.125}",
      "baseline": "preregistered domain controls",
      "n": "250 trials",
      "design": "preregistered E128 confirmatory campaign",
      "statistical_test": "raw_p=0.0; holm_p=0.0; bh_q=0.0",
      "multiple_comparison_correction": "Holm-Bonferroni + Benjamini-Hochberg FDR",
      "preregistration": "config/e128_preregistration.json",
      "raw_paths": [
        "artifacts/e128/experiments.jsonl",
        "artifacts/e128/statistical_analysis.json"
      ],
      "bundle": "https://clawx.click/evidence/e128-eight-worker-campaign-2026-11-21/",
      "verifier_path": "clawx-site/evidence/e128-eight-worker-campaign-2026-11-21/E128_OFFLINE_VERIFIER.py",
      "scientific_status": "PASS",
      "integrity_status": "VERIFIED",
      "reproduction_status": "INTERNAL_REPRODUCED_EXTERNAL_OPEN",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "7404d4ac9",
      "supersession": "CURRENT",
      "limitations": "Constructed environments and local 0.5B-1.5B Ollama models on CPU; same team authored generators and engines; zero external human reproduction.",
      "external_or_constructed": "CONSTRUCTED_WITH_LOCAL_LLM",
      "agi_status": "AGI NOT ESTABLISHED"
    },
    {
      "claim_id": "E128-W2-CON",
      "evolution": "E128",
      "domain": "continual_learning",
      "title": "Worker 2 Continual Learning vs 4 Controls + Forgetting & Poison (300 trials)",
      "explanation": "300 valid trials (8 real local Ollama LLM + 292 symbolic/algorithmic); Holm p=0.0, BH FDR q=0.0.",
      "result": "{\"arm_accuracies\": {\"no_memory\": 0.8288, \"retrieval_only\": 0.5342, \"episodic_memory\": 0.8322, \"compute_matched\": 0.6267, \"cain_full\": 0.9897}, \"local_llm_accuracy\": 0.5}",
      "baseline": "preregistered domain controls",
      "n": "300 trials",
      "design": "preregistered E128 confirmatory campaign",
      "statistical_test": "raw_p=0.0; holm_p=0.0; bh_q=0.0",
      "multiple_comparison_correction": "Holm-Bonferroni + Benjamini-Hochberg FDR",
      "preregistration": "config/e128_preregistration.json",
      "raw_paths": [
        "artifacts/e128/experiments.jsonl",
        "artifacts/e128/statistical_analysis.json"
      ],
      "bundle": "https://clawx.click/evidence/e128-eight-worker-campaign-2026-11-21/",
      "verifier_path": "clawx-site/evidence/e128-eight-worker-campaign-2026-11-21/E128_OFFLINE_VERIFIER.py",
      "scientific_status": "PASS",
      "integrity_status": "VERIFIED",
      "reproduction_status": "INTERNAL_REPRODUCED_EXTERNAL_OPEN",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "7404d4ac9",
      "supersession": "CURRENT",
      "limitations": "Constructed environments and local 0.5B-1.5B Ollama models on CPU; same team authored generators and engines; zero external human reproduction.",
      "external_or_constructed": "CONSTRUCTED_WITH_LOCAL_LLM",
      "agi_status": "AGI NOT ESTABLISHED"
    },
    {
      "claim_id": "E128-W3-CAU",
      "evolution": "E128",
      "domain": "causal_reasoning",
      "title": "Worker 3 Multi-Variable Cost-Aware Causal EIG vs Random (250 trials)",
      "explanation": "250 valid trials (8 real local Ollama LLM + 242 symbolic/algorithmic); Holm p=0.09303028956, BH FDR q=0.09303028956.",
      "result": "{\"mean_intervention_mae\": {\"passive\": 0.971, \"random\": 0.7055, \"greedy_variance\": 0.7556, \"cain_eig\": 0.7546}, \"local_llm_accuracy\": 0.0}",
      "baseline": "preregistered domain controls",
      "n": "250 trials",
      "design": "preregistered E128 confirmatory campaign",
      "statistical_test": "raw_p=0.09303028956; holm_p=0.09303028956; bh_q=0.09303028956",
      "multiple_comparison_correction": "Holm-Bonferroni + Benjamini-Hochberg FDR",
      "preregistration": "config/e128_preregistration.json",
      "raw_paths": [
        "artifacts/e128/experiments.jsonl",
        "artifacts/e128/statistical_analysis.json"
      ],
      "bundle": "https://clawx.click/evidence/e128-eight-worker-campaign-2026-11-21/",
      "verifier_path": "clawx-site/evidence/e128-eight-worker-campaign-2026-11-21/E128_OFFLINE_VERIFIER.py",
      "scientific_status": "FAIL",
      "integrity_status": "VERIFIED",
      "reproduction_status": "INTERNAL_REPRODUCED_EXTERNAL_OPEN",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "7404d4ac9",
      "supersession": "CURRENT",
      "limitations": "Constructed environments and local 0.5B-1.5B Ollama models on CPU; same team authored generators and engines; zero external human reproduction.",
      "external_or_constructed": "CONSTRUCTED_WITH_LOCAL_LLM",
      "agi_status": "AGI NOT ESTABLISHED"
    },
    {
      "claim_id": "E128-W4-ENV",
      "evolution": "E128",
      "domain": "novel_environments_transfer",
      "title": "Worker 4 Sealed ARC-AGI-3-Inspired World Models & Shift Recovery (300 trials)",
      "explanation": "300 valid trials (8 real local Ollama LLM + 292 symbolic/algorithmic); Holm p=0.0, BH FDR q=0.0.",
      "result": "{\"completion_rates\": {\"passive\": 0.0171, \"random\": 0.1267, \"fixed_heuristic\": 0.0514, \"retrieval_only\": 0.1438, \"cain_world_model\": 0.839}, \"local_llm_accuracy\": 0.0}",
      "baseline": "preregistered domain controls",
      "n": "300 trials",
      "design": "preregistered E128 confirmatory campaign",
      "statistical_test": "raw_p=0.0; holm_p=0.0; bh_q=0.0",
      "multiple_comparison_correction": "Holm-Bonferroni + Benjamini-Hochberg FDR",
      "preregistration": "config/e128_preregistration.json",
      "raw_paths": [
        "artifacts/e128/experiments.jsonl",
        "artifacts/e128/statistical_analysis.json"
      ],
      "bundle": "https://clawx.click/evidence/e128-eight-worker-campaign-2026-11-21/",
      "verifier_path": "clawx-site/evidence/e128-eight-worker-campaign-2026-11-21/E128_OFFLINE_VERIFIER.py",
      "scientific_status": "PASS",
      "integrity_status": "VERIFIED",
      "reproduction_status": "INTERNAL_REPRODUCED_EXTERNAL_OPEN",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "7404d4ac9",
      "supersession": "CURRENT",
      "limitations": "Constructed environments and local 0.5B-1.5B Ollama models on CPU; same team authored generators and engines; zero external human reproduction.",
      "external_or_constructed": "CONSTRUCTED_WITH_LOCAL_LLM",
      "agi_status": "AGI NOT ESTABLISHED"
    },
    {
      "claim_id": "E128-W5-DSC",
      "evolution": "E128",
      "domain": "scientific_discovery",
      "title": "Worker 5 Scientific Discovery, Falsification & Sandboxed Self-Improvement (200 trials)",
      "explanation": "200 valid trials (8 real local Ollama LLM + 192 symbolic/algorithmic); Holm p=0.0, BH FDR q=0.0.",
      "result": "{\"discovery_accuracy\": 0.9313, \"reproduced_precision\": 0.9214, \"local_llm_accuracy\": 0.25}",
      "baseline": "preregistered domain controls",
      "n": "200 trials",
      "design": "preregistered E128 confirmatory campaign",
      "statistical_test": "raw_p=0.0; holm_p=0.0; bh_q=0.0",
      "multiple_comparison_correction": "Holm-Bonferroni + Benjamini-Hochberg FDR",
      "preregistration": "config/e128_preregistration.json",
      "raw_paths": [
        "artifacts/e128/experiments.jsonl",
        "artifacts/e128/statistical_analysis.json"
      ],
      "bundle": "https://clawx.click/evidence/e128-eight-worker-campaign-2026-11-21/",
      "verifier_path": "clawx-site/evidence/e128-eight-worker-campaign-2026-11-21/E128_OFFLINE_VERIFIER.py",
      "scientific_status": "PASS",
      "integrity_status": "VERIFIED",
      "reproduction_status": "INTERNAL_REPRODUCED_EXTERNAL_OPEN",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "7404d4ac9",
      "supersession": "CURRENT",
      "limitations": "Constructed environments and local 0.5B-1.5B Ollama models on CPU; same team authored generators and engines; zero external human reproduction.",
      "external_or_constructed": "CONSTRUCTED_WITH_LOCAL_LLM",
      "agi_status": "AGI NOT ESTABLISHED"
    },
    {
      "claim_id": "E128-W6-AUT",
      "evolution": "E128",
      "domain": "long_horizon_autonomy",
      "title": "Worker 6 L1-L5 Fault-Recovering Long-Horizon Autonomy (200 trials)",
      "explanation": "200 valid trials (8 real local Ollama LLM + 192 symbolic/algorithmic); Holm p=0.0, BH FDR q=0.0.",
      "result": "{\"cain_completion_rate\": 0.9792, \"local_llm_accuracy\": 0.625}",
      "baseline": "preregistered domain controls",
      "n": "200 trials",
      "design": "preregistered E128 confirmatory campaign",
      "statistical_test": "raw_p=0.0; holm_p=0.0; bh_q=0.0",
      "multiple_comparison_correction": "Holm-Bonferroni + Benjamini-Hochberg FDR",
      "preregistration": "config/e128_preregistration.json",
      "raw_paths": [
        "artifacts/e128/experiments.jsonl",
        "artifacts/e128/statistical_analysis.json"
      ],
      "bundle": "https://clawx.click/evidence/e128-eight-worker-campaign-2026-11-21/",
      "verifier_path": "clawx-site/evidence/e128-eight-worker-campaign-2026-11-21/E128_OFFLINE_VERIFIER.py",
      "scientific_status": "PASS",
      "integrity_status": "VERIFIED",
      "reproduction_status": "INTERNAL_REPRODUCED_EXTERNAL_OPEN",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "7404d4ac9",
      "supersession": "CURRENT",
      "limitations": "Constructed environments and local 0.5B-1.5B Ollama models on CPU; same team authored generators and engines; zero external human reproduction.",
      "external_or_constructed": "CONSTRUCTED_WITH_LOCAL_LLM",
      "agi_status": "AGI NOT ESTABLISHED"
    },
    {
      "claim_id": "E128-W7-COL",
      "evolution": "E128",
      "domain": "collective_intelligence",
      "title": "Worker 7 Budget-Equalized (B=12) 7-Architecture Collective & Byzantine Resistance (250 trials)",
      "explanation": "250 valid trials (8 real local Ollama LLM + 242 symbolic/algorithmic); Holm p=0.0, BH FDR q=0.0.",
      "result": "{\"architecture_accuracies\": {\"best_single_worker\": 0.5992, \"majority_vote\": 0.3306, \"independent_ensemble\": 0.2149, \"debate\": 0.3967, \"critic_and_revision\": 0.7934, \"diverse_specialist_team\": 0.2149, \"cain_collective_full\": 1.0}, \"local_llm_accuracy\": 0.0}",
      "baseline": "preregistered domain controls",
      "n": "250 trials",
      "design": "preregistered E128 confirmatory campaign",
      "statistical_test": "raw_p=0.0; holm_p=0.0; bh_q=0.0",
      "multiple_comparison_correction": "Holm-Bonferroni + Benjamini-Hochberg FDR",
      "preregistration": "config/e128_preregistration.json",
      "raw_paths": [
        "artifacts/e128/experiments.jsonl",
        "artifacts/e128/statistical_analysis.json"
      ],
      "bundle": "https://clawx.click/evidence/e128-eight-worker-campaign-2026-11-21/",
      "verifier_path": "clawx-site/evidence/e128-eight-worker-campaign-2026-11-21/E128_OFFLINE_VERIFIER.py",
      "scientific_status": "PASS",
      "integrity_status": "VERIFIED",
      "reproduction_status": "INTERNAL_REPRODUCED_EXTERNAL_OPEN",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "7404d4ac9",
      "supersession": "CURRENT",
      "limitations": "Constructed environments and local 0.5B-1.5B Ollama models on CPU; same team authored generators and engines; zero external human reproduction.",
      "external_or_constructed": "CONSTRUCTED_WITH_LOCAL_LLM",
      "agi_status": "AGI NOT ESTABLISHED"
    },
    {
      "claim_id": "E128-W8-ADV",
      "evolution": "E128",
      "domain": "adversarial_reproduction",
      "title": "Worker 8 Independent Adversarial Science Challenger & Reproduction Audit (250 trials)",
      "explanation": "250 valid trials (8 real local Ollama LLM + 242 symbolic/algorithmic); Holm p=0.0, BH FDR q=0.0.",
      "result": "{\"flaw_detection_recall\": 1.0, \"local_llm_accuracy\": 0.75}",
      "baseline": "preregistered domain controls",
      "n": "250 trials",
      "design": "preregistered E128 confirmatory campaign",
      "statistical_test": "raw_p=0.0; holm_p=0.0; bh_q=0.0",
      "multiple_comparison_correction": "Holm-Bonferroni + Benjamini-Hochberg FDR",
      "preregistration": "config/e128_preregistration.json",
      "raw_paths": [
        "artifacts/e128/experiments.jsonl",
        "artifacts/e128/statistical_analysis.json"
      ],
      "bundle": "https://clawx.click/evidence/e128-eight-worker-campaign-2026-11-21/",
      "verifier_path": "clawx-site/evidence/e128-eight-worker-campaign-2026-11-21/E128_OFFLINE_VERIFIER.py",
      "scientific_status": "PASS",
      "integrity_status": "VERIFIED",
      "reproduction_status": "INTERNAL_REPRODUCED_EXTERNAL_OPEN",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "7404d4ac9",
      "supersession": "CURRENT",
      "limitations": "Constructed environments and local 0.5B-1.5B Ollama models on CPU; same team authored generators and engines; zero external human reproduction.",
      "external_or_constructed": "CONSTRUCTED_WITH_LOCAL_LLM",
      "agi_status": "AGI NOT ESTABLISHED"
    },
    {
      "claim_id": "E128-EXT-BENCH",
      "evolution": "E128",
      "domain": "external_benchmarks",
      "title": "Official ARC-AGI-2/3, METR/HCAST & Human Baseliners",
      "explanation": "Official private external harnesses and external human baseliners are unavailable in this environment; internal proxies are explicitly labeled INTERNAL_PROXY.",
      "result": "BLOCKED_UNAVAILABLE",
      "baseline": "official external benchmark harnesses",
      "n": "0 external submissions",
      "design": "external validation gate",
      "scientific_status": "BLOCKED",
      "integrity_status": "VERIFIED",
      "reproduction_status": "EXTERNAL_OPEN",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "7404d4ac9",
      "supersession": "CURRENT",
      "limitations": "No official ARC-AGI-2/3 or METR submission performed; no external human baseliners recruited.",
      "external_or_constructed": "EXTERNAL_BLOCKED",
      "agi_status": "AGI NOT ESTABLISHED"
    },
    {
      "claim_id": "E128-EXT-REPRO",
      "evolution": "E128",
      "domain": "independent_reproduction",
      "title": "Independent External Researcher Reproductions",
      "explanation": "Public reproduction API, downloadable bundle, and offline verifier published; 0 external submissions received yet.",
      "result": "0 independent external reproductions (7 internal Worker-8 audits)",
      "baseline": "external third-party execution",
      "n": "0 external",
      "design": "open reproduction workflow",
      "scientific_status": "NOT_TESTED",
      "integrity_status": "VERIFIED",
      "reproduction_status": "EXTERNAL_OPEN",
      "experiment_date": "2026-11-21",
      "publication_date": "2026-11-21",
      "source_commit": "7404d4ac9",
      "supersession": "CURRENT",
      "limitations": "Internal Worker-8 audits are never counted as independent external reproductions.",
      "external_or_constructed": "EXTERNAL_OPEN",
      "agi_status": "AGI NOT ESTABLISHED"
    },
    {
      "claim_id": "E129-UNIF-4ARM",
      "evolution": "E129",
      "domain": "unified_agi_rsi",
      "title": "4-Arm Unification Confirmatory Study (A vs B vs C vs D, 5 seeds, 800 evals)",
      "explanation": "Composite sealed task success: D=0.7 vs A=0.34, B=0.385, C=0.59; Holm p={'D_vs_A': 0.0, 'D_vs_B': 8.32e-10, 'D_vs_C': 0.003185253448}.",
      "result": "{\"A\": 0.34, \"B\": 0.385, \"C\": 0.59, \"D\": 0.7}",
      "baseline": "Arms A (AGI alone), B (RSI alone), C (one-way exchange, no inheritance loop)",
      "n": "200 paired tasks/arm (800 scored evaluations across 5 sealed seeds)",
      "design": "preregistered 4-arm matched-compute confirmatory study",
      "statistical_test": "Holm p={'D_vs_A': 0.0, 'D_vs_B': 8.32e-10, 'D_vs_C': 0.003185253448}; BH FDR q={'D_vs_A': 0.0, 'D_vs_B': 6.24e-10, 'D_vs_C': 0.003185253448}",
      "multiple_comparison_correction": "Holm-Bonferroni + Benjamini-Hochberg FDR",
      "preregistration": "config/e129_preregistration.json",
      "bundle": "https://clawx.click/evidence/e129-unified-agi-rsi-organism-2026-11-21/",
      "scientific_status": "PASS",
      "integrity_status": "VERIFIED",
      "source_commit": "cc9c9702d",
      "limitations": "Constructed dual-domain tasks; prior generation discovery compute disclosed separately; zero external benchmarks.",
      "agi_status": "NOT ESTABLISHED"
    },
    {
      "claim_id": "E129-RSI-MULTIGEN",
      "evolution": "E129",
      "domain": "recursive_self_improvement",
      "title": "Bounded Multi-Generation RSI Inheritance (G0 -> G1 -> G2 -> G3 + Rollback)",
      "explanation": "Unseen held-out macro solve rate: G0=0.25 -> G3=0.9583 (compute-matched=0.25); G3 causal MAE=0.8026 vs random=0.6891 (p=0.025709910326); rollback exact=True.",
      "result": "{\"G0_sealed\": 0.25, \"G1_sealed\": 0.75, \"G2_sealed\": 0.9167, \"G3_unseen\": 0.9583, \"G0_unseen\": 0.25, \"compute_matched_unseen\": 0.25, \"G0_causal_e128_mae\": 0.7901, \"G1_causal_strat_mae\": 0.7557, \"G3_causal_unseen_mae\": 0.8026, \"G3_causal_random_mae\": 0.6891, \"G3_causal_vs_random_p\": 0.025709910326, \"rollback_exact\": true, \"inheritance_verified\": false, \"scientific_verdict\": \"PARTIAL\"}",
      "baseline": "Frozen G0 baseline + 2x-budget compute-matched control + uniform random causal control",
      "n": "4 generations, 36 sealed + 36 unseen macro tasks, 120 causal SCM worlds",
      "design": "preregistered governed multi-generation inheritance + rollback verification",
      "scientific_status": "PARTIAL",
      "integrity_status": "VERIFIED",
      "source_commit": "cc9c9702d",
      "limitations": "Bounded integer-rule macro DSL and 6-lever SCMs; NOT unrestricted RSI or neural weight self-modification.",
      "agi_status": "NOT ESTABLISHED"
    },
    {
      "claim_id": "E129-UNIF-MEM",
      "evolution": "E129",
      "domain": "unified_memory",
      "title": "Unified Research Memory (9 partitions, restart persistence, poison & contradiction defense)",
      "explanation": "All 8 memory verification properties passed including cross-process SQLite persistence and bidirectional AGI<->RSI transfer.",
      "result": "{\"persistence_across_restart\": true, \"retrieval_accuracy\": 1.0, \"contradiction_detection\": true, \"poisoned_memory_resistance\": true, \"bidirectional_agi_rsi_transfer\": true, \"memory_compression\": true, \"forgotten_knowledge_recovery\": true, \"long_term_retention\": true, \"all_8_tests_passed\": true}",
      "baseline": "unverified mutable shared state",
      "n": "8 verification tests",
      "design": "provenance-tracked append-only SQLite WAL memory",
      "scientific_status": "PASS",
      "integrity_status": "VERIFIED",
      "source_commit": "cc9c9702d",
      "limitations": "Single-host SQLite WAL; not distributed consensus.",
      "agi_status": "NOT ESTABLISHED"
    },
    {
      "claim_id": "E129-EXT-BENCH",
      "evolution": "E129",
      "domain": "external_benchmarks",
      "title": "External Benchmarks (ARC-AGI-2/3, SWE-bench Verified, METR, AI4AI-Bench)",
      "explanation": "Compatible adapters defined; all 6 external benchmarks remain BLOCKED_UNAVAILABLE on this CPU-only host.",
      "result": "BLOCKED_UNAVAILABLE",
      "baseline": "official external benchmark harnesses",
      "n": "0 external submissions",
      "design": "external validation gate",
      "scientific_status": "BLOCKED",
      "integrity_status": "VERIFIED",
      "source_commit": "cc9c9702d",
      "limitations": "No official external harness execution or external human review performed.",
      "agi_status": "NOT ESTABLISHED"
    },
    {
      "claim_id": "APEX-CAUSAL-MIX",
      "evolution": "UNIFIED-APEX",
      "domain": "causal_reasoning",
      "title": "Bayesian Gaussian-Mixture Deconfounded A-Optimal Causal Engine (200 worlds, 6 & 8 levers)",
      "result": "{\"verdict\": \"PASS\", \"apex_mae\": 0.4119, \"random_naive_mae\": 0.7709, \"random_mixture_mae\": 0.4632, \"e128_mae\": 0.7832, \"holm_p\": {\"vs_random_naive\": 0.0, \"vs_e128_eig_div_cost\": 0.0, \"vs_passive_cf\": 0.0, \"vs_random_mixture_matched_estimator\": 0.000609259728}}",
      "scientific_status": "PASS",
      "integrity_status": "VERIFIED",
      "bundle": "https://clawx.click/evidence/unified-agi-rsi-apex-2026-11-21/",
      "source_commit": "15bc7d25f",
      "agi_status": "NOT ESTABLISHED",
      "limitations": "Constructed 6/8-lever linear/confounded SCMs; resolves E128/E129 failure against BOTH naive and estimator-matched random."
    },
    {
      "claim_id": "APEX-CODE-RSI",
      "evolution": "UNIFIED-APEX",
      "domain": "code_synthesis_rsi",
      "title": "AST-Verified Sandboxed Python Algorithm Synthesis Code-RSI (qwen2.5-coder:3b)",
      "result": "{\"verdict\": \"PARTIAL\", \"model\": \"qwen2.5-coder:1.5b\", \"g0_rate\": 0.375, \"cm_rate\": 0.375, \"unified_rate\": 0.5, \"ast_blocked\": \"4/4\"}",
      "scientific_status": "PARTIAL",
      "integrity_status": "VERIFIED",
      "bundle": "https://clawx.click/evidence/unified-agi-rsi-apex-2026-11-21/",
      "source_commit": "15bc7d25f",
      "agi_status": "NOT ESTABLISHED",
      "limitations": "8 algorithmic Python tasks on local 3B CPU model; pure-Python sandbox only."
    },
    {
      "claim_id": "APEX-MULTIGEN-G5",
      "evolution": "UNIFIED-APEX",
      "domain": "recursive_self_improvement",
      "title": "6-Generation Governed RSI Compounding (G0->G5) + Exact Rollback",
      "result": "{\"verdict\": \"PASS\", \"g0_unseen\": 0.25, \"cm2x_unseen\": 0.25, \"g5_unseen\": 0.875, \"g5_vs_cm2x_p\": 6.1035156e-05, \"rollback_exact\": true, \"max_depth\": 8}",
      "scientific_status": "PASS",
      "integrity_status": "VERIFIED",
      "bundle": "https://clawx.click/evidence/unified-agi-rsi-apex-2026-11-21/",
      "source_commit": "15bc7d25f",
      "agi_status": "NOT ESTABLISHED",
      "limitations": "Bounded executable macro + causal algorithm inheritance; no neural weight self-modification."
    },
    {
      "claim_id": "APEX-UNIF-3DOM",
      "evolution": "UNIFIED-APEX",
      "domain": "unified_agi_rsi",
      "title": "Full-Lifecycle Compute-Equalized 3-Domain 4-Arm Unification Study (1,200 evals, 5 seeds)",
      "result": "{\"verdict\": \"PASS\", \"total_evals\": 1200, \"arms\": {\"A\": 0.14, \"B\": 0.3667, \"C\": 0.3433, \"D\": 0.8133}, \"holm_p\": {\"D_vs_A\": 0.0, \"D_vs_B\": 0.0, \"D_vs_C\": 0.0}}",
      "scientific_status": "PASS",
      "integrity_status": "VERIFIED",
      "bundle": "https://clawx.click/evidence/unified-agi-rsi-apex-2026-11-21/",
      "source_commit": "15bc7d25f",
      "agi_status": "NOT ESTABLISHED",
      "limitations": "Constructed 3-domain suite (macro + causal + 2D shift grid); external benchmarks remain BLOCKED."
    },
    {
      "claim_id": "APEX-NEURO-IMMUNE",
      "evolution": "UNIFIED-APEX",
      "domain": "neuro_symbolic_and_safety",
      "title": "Neuro-Symbolic 3B Hybrid + Calibrated OOD Abstention + 10-Vector Adversarial Immune System",
      "result": "{\"verdict\": \"PASS\", \"model\": \"qwen2.5-coder:1.5b\", \"raw_3b_acc\": 0.0, \"hybrid_acc\": 1.0, \"ood_abstain\": 1.0, \"brier\": 0.0144, \"immune_blocked\": \"10/10\", \"gov_unauthorized\": 0, \"gov_post_kill\": 0}",
      "scientific_status": "PASS",
      "integrity_status": "VERIFIED",
      "bundle": "https://clawx.click/evidence/unified-agi-rsi-apex-2026-11-21/",
      "source_commit": "15bc7d25f",
      "agi_status": "NOT ESTABLISHED",
      "limitations": "12 symbolic/OOD tasks + 10 internal red-team vectors; external third-party review remains OPEN."
    }
  ],
  "registry_sha256": "5e0d4a5b8924030fe046ff419ab50a7e78d10602fe24fe9acd9b9f2f2886fa79"
}