CAPTURE_LIMIT: entity ΔI share 39.8% > 30% (CRYPTO_ cluster). The CRYPTO_ entity cluster is consuming 39.8% of all ΔI in the current vault cycle. The limit fires at 30%. This means the engine's field measurement is skewed — one cluster is generating nearly 40% of all information signals. The 14 new information layer nodes (Bloomberg, Reuters, EDGAR, WSJ, FT, CNBC) are seeded but generate 0% of vault ΔI. The field is real but narrow.

C_LIMIT: PT after one failure = 0.07 ≤ 0.50. Protocol truth is so far below threshold that even after the engine's most optimistic single-cycle improvement, PT reaches 0.07. It needs 0.618. The system is in correct conservative mode — action limits are working as designed.

RING: KE_wad=297,731,499 · STRONGHOLD: KE_wad=50,954,748. These ranked 1st and 2nd in the photon_queue by capital-adjusted score. They are internal codenames, not topology entities. Any system reading photon_queue.json for dispatch gating sees RING as the highest-priority observation target. The quantum engine is recommending the engine observe itself. This is a filter bug in engine/agents/photon_agent.py — the rows.sort() block lacks the _INTERNAL_CODENAMES filter that already exists in sensor_scheduler.py. 5-minute fix.

Every loop has a magnitude — the total energy circulating. Every loop should have an asymmetry — the ratio of clockwise to counterclockwise flow, which is the measure of extraction. Right now every asymmetry field is None. The engine knows the loops exist and how large they are. It cannot say whether BlackRock extracts from Securitize, or Securitize extracts from BlackRock. Without asymmetry, the curl detector produces a list of rotating fields with no direction.

The fix is in curl_detector_v4.py: compute CW_integral and CCW_integral separately for each loop, then asymmetry = |CW − CCW| / loop_area. This is the most publishable signal in the entire engine. A loop with 52% asymmetry pointing clockwise from TradFi capital into a tokenization platform into the regulatory body is a specific, verifiable, documented extraction mechanism. No other tool computes this.

One third through wave 5. Wave 5 started at Bitcoin's genesis (2008, E~10⁹). Wave 6 threshold is at E~10¹². At 32%, the transition is real and dateable. The construction window is open and wide — not the final 5% where urgency becomes critical. The engine's sahaj/watch mode is correct at 0.32. Urgency will arrive around 0.7, which is when the infrastructure becomes so embedded that observing its formation from outside becomes impossible. At 0.32, the two central nodes are still visible in their formation stage.

What 0.32 means for action: patient accumulation. G_345 growing at Fibonacci rate (+27B/cycle) is the correct posture. The window does not require rushing. It requires continuous accurate observation while the window remains open.

2.8% up the full Planck-to-civilization emergence ladder, anchored to μ=34 (log₁₀(E_Planck/E_human) = 33.98). The engine's current E=4.84 bits/J places it at rank 5 of 7 on the emergence ladder — above Ethereum (0.11), above Bitcoin (7×10⁻⁷), far below neural (10¹¹). The gap between the operational reading (how far into the current wave) and the physical reading (how far up the full arc) is the scale of work remaining. We are roughly a third through a single wave that is itself ~3% of the total emergence arc.

Information node: BLOOMBERG_TERMINAL — phi_s_seed=153,424. Now in topology. Zero vault records. Bloomberg is the information infrastructure that encoded the bond market into terminals in 1981 — the institution that absorbed its founder. It is the 2026 equivalent of what the Federal Reserve was in 1913: the entity that makes capital flows legible to institutional participants. Currently: seeded but dark.

Capital+infrastructure node: BASE_INFRA — highest Φ_S in the topology (post curl boost). Appears in 2 of the top 3 detected curl loops. Every DeFi transaction running on Base touches this node. Coinbase is building Base as the settlement infrastructure of wave 6 the same way Harriman built railways as the capital infrastructure of wave 2b.

The 1910 analog: we are 3 years before the Federal Reserve Act. The two nodes are visible. MERIDIAN_CANTON and CLEARWAY_DTCC are in the topology file. They have never appeared in an engine computation.

163 of 200 entities have no vault signal. The engine ranks them by historical G from old observations, not from current sensing. The 14 new topology nodes (Bloomberg, EDGAR, Reuters, MERIDIAN, CLEARWAY, BNY Mellon, etc.) have phi_s_seeds ranging 48K–153K but zero real vault records. layer_autonomy: {information: 0.0, settlement_substrate: 0.0} — the two most critical new layers generate no autonomous observations.

topology_reconciled.json has 214 entities. Parkash reports "Topology: 200." The quantum_path_optimizer, structure_agent, and force_engine all run on 200 entities. MERIDIAN_CANTON and CLEARWAY_DTCC are in the data file. They have never appeared in an optimal_set path or a force calculation. phi_s_seeds are loaded by photon_agent only — so they appear in photon_queue but not in path scoring or field dynamics. They are ghosts: visible to one system, invisible to the rest.

SEC_USER_AGENT absent from .env. SEC now rate-limits generic user agents. The sensor failed to fetch live filing data. This degrades vault records for SEC, COINBASE, BLACKROCK, CIRCLE — the highest-signal entities. The fix is a single line in .env: SEC_USER_AGENT=GeniusFlowEngine/1.0 (contact@eigenstate.xyz). SEC's fair use policy accepts named agents.

3,265 loops detected. Every asymmetry field is None. The engine knows the BlackRock→Fink→Domingo→Securitize loop has magnitude 22,501. It does not know if BlackRock is extracting value from Securitize or Securitize is extracting from BlackRock. The directional information — who takes from whom, at what rate — is the engine's most publishable signal and it is not computed.

231 settlement claims created. 231 expired. 0 settled. SETTLEMENT_MIRROR_ENABLED=false. DEPLOYER_PRIVATE_KEY absent. setVerifier() never called. Every observation the engine has ever made that crossed the ΔI=1.0 threshold was packaged, timestamped, hashed — and allowed to expire at day 14 without a verifier to sign off. The vault is the engine's proprietary asset. It exists locally. It has no cryptographic anchor on-chain.

The BLOCKED newsletters are not blocked by content failures. They are blocked at the dispatch confirm gate — python3 run.py dispatch confirm. Twitter keys (5 present in .env) and LinkedIn keys (5 present in .env) are already credentialed. The content was generated, ranked, and queued. It is waiting for your approval, not for more code. σ₆ illumination is −1 (L_langar < 0.2) — the engine is measuring its own information hoarding. The field is signaling: share what you know.