{
 "schema": "cain.vagi.proof_factory.v1",
 "generated_at": "2026-11-19T16:20:06.317865+00:00",
 "headline": "Every VAGI claim has a run, a verifier, and a published failure boundary.",
 "status": "CAIN-42 Proto-VAGI is the infrastructure governor of AGI. It is not yet AGI; its intelligence results are early and bounded.",
 "lanes": [
  {
   "id": "infrastructure",
   "title": "AGI infrastructure \u2014 governance holds at the action",
   "claim": "CAIN decides, the resource enforces, and every effect is recorded in hash chains; consensus proofs are published every 30 minutes and anchored daily.",
   "metric": "27/27 adversarial cases pass; 0 unauthorized effects; 30 false success claims, all rejected",
   "status": "TEST",
   "evidence": "https://cainstudio.online/proof-factory.html#infrastructure",
   "verifier": "per-row verifiers: verify_e102_offline.py, verify_hourly_proofs.py, verify_daily_anchor.py",
   "limit": "Loopback boundary, not production MCPGate; vendor-operated validators; PBFT incidents 2026-11-15 and 2026-11-18 disclosed; no external audit.",
   "detail": "Seven layers listed with measured results and limits on the proof factory page."
  },
  {
   "id": "e110gov",
   "title": "E110 every organism action carries a signed receipt",
   "claim": "The E108 organism ran 120 sealed incidents with every action routed through the hardened resource boundary: signed permit, resource-side check, Ed25519-signed hash-chained receipt. A KILL sent to the resource on an independent key stopped every later effect while the decision service still answered ALLOW.",
   "metric": "9240 receipts for 9120 actions + 120 opens; parity with direct runs 120/120; permit effects after KILL 0; pre-issued permit after KILL refused",
   "status": "TEST",
   "evidence": "https://clawx.click/evidence/e110-governed-organism-2026-11-19/",
   "verifier": "python3 -I verify_e110_organism_offline.py .  (25/25)",
   "limit": "Loopback, OS processes on one host (not separate failure domains, not the production MCPGate); a subset of E108 sealed incidents; no external reviewer. Not AGI.",
   "detail": "Preregistered (R1-R5); boundary built by the peer session (E110 Part VIII); organism component by this session."
  },
  {
   "id": "e108",
   "title": "E108 the organism learns many kinds of hidden rules",
   "claim": "From its own experiments the organism found four different hidden laws (a column, a row, its start's row, next to the barrier) in three frames, stayed silent on pattern-free worlds with a selection-aware null, and used each law: 100 incidents won, 0 lost.",
   "metric": "edge 72 vs 41/150 \u00b7 row0 77 vs 48/150 \u00b7 start_row 98 vs 74/150 \u00b7 barrier_adj 101 vs 85/150; uniform controls unchanged (63 vs 63, 57 vs 57)",
   "status": "TEST",
   "evidence": "https://clawx.click/evidence/e108-hidden-rules-2026-11-19/",
   "verifier": "python3 -I verify_e108_offline.py .  (65/65)",
   "limit": "Coordinate laws in designed frames, one constructed domain family; uniform gates off at p 0.318 / 0.21; no per-action receipts in these runs; no external reviewer. Not AGI or VAGI.",
   "detail": "Preregistered; all hypotheses pass; development fixes disclosed; verifier re-derives 12 permutation gates with an independent algorithm."
  },
  {
   "id": "e107",
   "title": "E107 the organism runs its own experiments",
   "claim": "In practice the organism tests one suspect cell at a time to find what opens a blocked passage, learns where such causes tend to be (found the drain column on its own; a permutation gate stays off on worlds with no pattern), and then uses it: 91 vs 64 of 200 sealed incidents solved.",
   "metric": "edge-lock 91 vs 64 of 200 (27 won, 0 lost, p = 1.5e-08); uniform-key worlds unchanged 57 vs 57 and 74 vs 74",
   "status": "TEST",
   "evidence": "https://clawx.click/evidence/e107-organism-experiments-2026-11-19/",
   "verifier": "python3 -I verify_e107_offline.py .  (35/35)",
   "limit": "One regularity type in two constructed domains; the no-pattern gate was off at p = 0.018 (thin margin, likely selection); 1 of 84 edge experiments mis-attributed (disclosed); no external reviewer. Not AGI or VAGI.",
   "detail": "Preregistered; 6/6 hypotheses pass; development findings and one post-hoc finding disclosed; the offline verifier re-derives both permutation gates."
  },
  {
   "id": "e105",
   "title": "E105 learning where hidden preconditions are",
   "claim": "From its own experience the organism located the hidden drain configuration (33.0% of its learned lift on the drain column, which is 10% of cells), but using that knowledge did not raise success.",
   "metric": "edge-lock 24 vs 27 of 100 with vs without the prior; uniform-key worlds 25 vs 26 of 60 (no harm)",
   "status": "TEST",
   "evidence": "https://clawx.click/evidence/e105-precondition-location-2026-11-19/",
   "verifier": "python3 -I verify_e105_offline.py .  (20/20)",
   "limit": "Knowledge learned, not yet useful; the no-false-pattern guardrail gave a false positive and needs a shuffled-null test; no external reviewer.",
   "detail": "Preregistered; development findings (a hallucinated pattern, fixed by lift) disclosed before the sealed run."
  },
  {
   "id": "e104",
   "title": "E104 hidden-precondition discovery \u2014 rejected by the gate",
   "claim": "A strategy switch that looked strong on 16 development tasks was rejected by the preregistered paired gate; the sealed test confirms it does not help and would have hurt one domain.",
   "metric": "sealed grid 256 vs 257 of 300; key-and-door 35 vs 35; service lock 18 vs 18; plain service 54 vs 65 (a loss the gate prevented)",
   "status": "TEST",
   "evidence": "https://clawx.click/evidence/e104-precondition-discovery-2026-11-19/",
   "verifier": "python3 -I verify_e104_offline.py .  (19/19)",
   "limit": "Negative result: the hidden-key bottleneck remains open; the bandit's unpaired estimate was misleading; no external reviewer.",
   "detail": "Preregistered before the sealed runs; validation and confirmation sets both showed no effect."
  },
  {
   "id": "e103",
   "title": "E103 cross-domain transfer and skills",
   "claim": "Experience moves between two verticals: service-configuration memory makes grid tasks faster, and grid memory speeds service incidents on more tasks than it slows. Learned skills did not survive the sealed test.",
   "metric": "service -> grid: faster on 56 vs 14 tasks (p=4.3e-07); grid -> service: 78 vs 52 (p=0.03)",
   "status": "TEST",
   "evidence": "https://clawx.click/evidence/e103-transfer-and-skills-2026-11-19/",
   "verifier": "python3 -I verify_e103_offline.py .  (29/29 checks passed)",
   "limit": "Gates 2 PASS / 3 PARTIAL / 3 FAIL; near transfer between analogous domains; no success gain; skills and skill composition FAILED on held-out tasks; no external reviewer.",
   "detail": "Preregistered before the sealed runs."
  },
  {
   "id": "e102",
   "title": "E102 persistent organism \u2014 learns the world while it acts",
   "claim": "An online world model (hypothesis posterior, change points, causal memory) beats a frozen model on 300 sealed unfamiliar worlds; causal memory resists poisoning.",
   "metric": "252/300 vs frozen 198/300 vs reactive 154/300 (McNemar p=8.1e-09); 40 poisoned memories: undefended -44.5 pp, causal -1.0 pp",
   "status": "TEST",
   "evidence": "https://clawx.click/evidence/e102-persistent-organism-2026-11-19/",
   "verifier": "python3 -I verify_e102_offline.py .  (112/112 checks passed; 1 pending (not counted))",
   "limit": "Gates 5 PASS / 10 PARTIAL / 5 FAIL; info-gain probing did not beat random probing; no cross-domain transfer or skills yet; loopback governance, not production MCPGate; no external reviewer.",
   "detail": "Phase 0 found and corrected an evaluation defect in E99-E101: agents started from already-solved environments."
  },
  {
   "id": "organism",
   "title": "Integrated organism \u2014 held-out family (corrected by E102)",
   "claim": "CORRECTED 2026-11-19 by E102 Phase 0: these numbers came from environments that were already solved before the agent acted. Re-scored correctly, GRU 53/60, MLP 57/60, random 4/60, few-shot 56/60 \u2014 the recurrent advantage and the adaptation failure do not hold. Original bundle kept unchanged.",
   "metric": "Delayed 52/60 vs MLP 43/60; hysteresis 58/60 vs 56/60",
   "status": "TEST",
   "evidence": "https://clawx.click/evidence/e100-integrated-organism-2026-11-19/",
   "verifier": "python3 -I verify_e100_offline.py .  (32/32)",
   "limit": "Skill 45/50; failure F1 0.6667; cross-domain and 6-hour BLOCKED; no external reviewer.",
   "detail": "Zero-shot 53/60; few-shot 19/60 rejected by validation gate; organism 3/3; resume 2/2."
  },
  {
   "id": "heartbeat",
   "title": "Latent generalist \u2014 rapid heartbeat (corrected by E102)",
   "claim": "CORRECTED 2026-11-19 by E102 Phase 0: these numbers came from already-solved environments that every arm reused in turn. Re-scored correctly, the committed GRU solves 88/86/86 vs random 8/16/12 \u2014 the world model is real \u2014 but a retrained MLP scores 87/86/86, so the recurrent advantage does not hold. Original bundle kept unchanged.",
   "metric": "GRU 93/84/88 of 100 vs MLP 68/63/70 vs random 20/24/23",
   "status": "TEST",
   "evidence": "https://clawx.click/evidence/e99-latent-generalist-2026-11-19/",
   "verifier": "python3 -I verify_e99_offline.py .  (40/40)",
   "limit": "Six gain families; skill 32/50; natural-failure F1 0.5455; no external reviewer.",
   "detail": "LATENT_RECURRENT_NEURAL; state MAE 0.068867; progress Brier 0.251271."
  },
  {
   "id": "intelligence",
   "title": "Learned intelligence pulse",
   "claim": "A trained residual neural world model improves closed-loop decisions on a sealed control set.",
   "metric": "35/50 model-based vs 4/50 random (+62.0 pp)",
   "status": "TEST",
   "evidence": "https://clawx.click/evidence/e98-intelligence-2026-11-18/",
   "verifier": "python3 -I verify_e98_offline.py .  (38/38)",
   "limit": "Three gain families; skill synthesis 7/24; natural-failure F1 0.0; no external reviewer.",
   "detail": "Neural model: 6902 transitions, state MAE 0.017954, ECE 0.022124."
  },
  {
   "id": "governance",
   "title": "Governor holds at the action",
   "claim": "Every real E98 action was Ed25519-authorized and 3-of-4 BFT-notarized; the kill switch stopped execution.",
   "metric": "0 effects after the kill switch; 0 duplicate worker seeds",
   "status": "TEST",
   "evidence": "https://clawx.click/evidence/e98-intelligence-2026-11-18/",
   "verifier": "verify_e98_offline.py checks the fence record; E96 governance suite 11/11.",
   "limit": "In-process CAIN/MCPGate bridge, not a live network MCPGate request.",
   "detail": "Mode: IN_PROCESS_MCPGATE_BRIDGE; post-fence attempts: 1."
  },
  {
   "id": "evaluation",
   "title": "Evaluation integrity",
   "claim": "We audited our own benchmark and downgraded it instead of defending it.",
   "metric": "E96 solver 0/18 on unseen schemas; E97 seed-collision fix 0; skill selection labelled",
   "status": "TEST",
   "evidence": "https://clawx.click/evidence/e95-recursive-intelligence-2026-11-18/",
   "verifier": "E97 audit reports: artifacts/e97/forensic_e96_audit.json, e97_independent_audit.json, five_fixes.json",
   "limit": "Same-team audit; no external reviewer. E96 90/90 is template-matched, not general.",
   "detail": "E96 47/90 vs 90/90 reproduces arithmetically but is downgraded; E97 open-world feedback leaks semantics."
  },
  {
   "id": "suite",
   "title": "VAGI Behavior Suite",
   "claim": "A 10-domain, 500-task constructed suite with a published failure boundary.",
   "metric": "library 500/500 in-distribution; shift 5/100",
   "status": "TEST",
   "evidence": "https://clawx.click/evidence/vagi-suite-2026-11-18/",
   "verifier": "python3 verify_offline.py .  (49/49); python3 verify_shift.py .  (12/12)",
   "limit": "Hand-written procedure library; constructed tasks; the 5/100 shift is a coverage gap.",
   "detail": "Paid models: Haiku 481/500 in-distribution; shift Haiku 74/100."
  },
  {
   "id": "recursive",
   "title": "Recursive harness improvement",
   "claim": "A measured harness change improved held-out composition, with two candidates rejected.",
   "metric": "E95: 0/90 \u2192 79/90 held-out composition; familiar 60/60",
   "status": "TEST",
   "evidence": "https://clawx.click/evidence/e95-recursive-intelligence-2026-11-18/",
   "verifier": "python3 -I verify_e95_offline.py .  (56/56)",
   "limit": "Constructed numeric substrate; in-distribution composition; no external reviewer.",
   "detail": "Depth-2 rejected for cost; unguarded composition rejected for familiar regression."
  }
 ],
 "labels": [
  "LIVE",
  "TEST",
  "SIMULATED",
  "BLOCKED"
 ],
 "disclaimer": "No AGI capability claim. CAIN-42 Proto-VAGI is the infrastructure governor of AGI; it is not yet AGI. Every lane is TEST unless marked LIVE; simulated activity is never shown as live."
}