{ "artifact_id": "neurobrick-system-review-2026-09-27", "observed_at": "2026-09-27T08:23:00+09:00", "inspection_window": "27 September 2026 · 08:19–08:23 KST", "scope": "Read-only inspection of the known operating fleet. Point-in-time internal disclosure, not live telemetry or independent attestation.", "inventory": { "compute_hosts": 7, "storage_appliances": 1, "execution_systems": 3, "mesh_registered": 9, "mesh_online": 7, "control_plane_containers": 7, "apple_silicon_installed_memory_gib": 224, "ci_vms_running": 3 }, "counting_boundary": "Seven compute hosts were accessed: the control host locally and six remote hosts over SSH. One NAS was observed through mounted shares on both Apple Silicon hosts. The mesh count is network identities: nine registered, seven online. It includes client devices and one registered machine that is not part of the operating fleet. The seven containers are the running database, cache, metrics and log services of the control stack on one host; the control-plane service itself was observed as a host process, and a separate second container stack on that host is not counted. Three running Linux CI guests were observed on three hosts; two further running Linux guests were not established as CI runners and are not counted. Containers, VMs and logical execution systems are not additional physical boxes. This is the known inspected fleet, not a discovery proof for every device.", "hosts": [ { "role": "Control & development", "count": 1, "evidence": "Local inspection: the daily access-gate measurement (06:00 KST) and its morning report (06:10 KST) completed with result 0; the primary-host observation task last ran 08:20 KST, result 0.", "boundary": "The local container engine was not running at inspection. Two collector tasks reported a nonzero last result. Task results do not grant release authority." }, { "role": "Research & primary execution", "count": 1, "evidence": "Remote runtime inspection: two trader processes, a domestic extended-hours trader and an order-enforcement process were present after the weekly restart at 03:43 KST (253 of 260 in-scope services restarted, six stopped as hung). The lineage index built at 07:10 KST.", "boundary": "Process presence does not prove every strategy is armed or orders have filled. The inspection fell on a weekend with the equity markets closed." }, { "role": "Local inference & Linux CI", "count": 1, "evidence": "Remote inspection: 128 GiB installed memory; two running Linux guests, one of them the CI guest (12 vCPU, 48 GiB).", "boundary": "Inference generation, model residency and distributed inference were not tested in this window." }, { "role": "Domestic rotation execution", "count": 1, "evidence": "Scheduled rotation task: last run 13:05 KST on 26 September, result 0; task Ready at inspection. The seven support containers of the control stack were running on this host.", "boundary": "Scheduled-task completion is not broker reconciliation and does not establish that the rotation lane is placing live orders." }, { "role": "Multi-account execution", "count": 1, "evidence": "Scheduled account worker: last run 08:22 KST, result 0; task Ready at inspection.", "boundary": "A worker result does not establish every account or strategy is eligible." }, { "role": "Fleet observation", "count": 1, "evidence": "Daily scan started 08:00 KST and was Running at inspection. Deadman watch: 07:00 KST, result 0. Deadman presence: 08:15 KST, result 0. Master-brain task: last run 20 September 18:30 KST, result 0.", "boundary": "The 08:00 scan had not completed inside the window. Scheduler records do not independently validate report contents or full fleet coverage." }, { "role": "Windows / WSL CI", "count": 1, "evidence": "Remote host reachable; one Linux CI guest Running; four self-hosted runner services Running; boot keepalive task Running.", "boundary": "The keepalive task carried a nonzero last-task result. Running runner services do not establish job success." } ], "systems": [ { "name": "NeuroPublic", "role": "Research, market data and primary execution", "detail": "A research and execution environment with domestic and US trader processes and separate order enforcement. Data readiness and strategy promotion are separate gates." }, { "name": "KR Rotation", "role": "Independent domestic execution", "detail": "A dedicated scheduled domestic rotation engine with its own execution and account boundary." }, { "name": "NeuroForAll", "role": "Multi-account execution", "detail": "An account-scoped execution platform with scheduled workers and separate account contexts." }, { "name": "Control Plane", "role": "Control and evidence integration", "detail": "A separate coordination layer supplying state, release controls and evidence services. It is not counted as a fourth independent execution system." } ], "capabilities": [ { "title": "Operate separate execution paths", "evidence": "Three execution systems across separate hosts", "detail": "Primary research and execution, domestic rotation and multi-account workers remain distinct.", "boundary": "An operating process or successful scheduled task does not prove a filled order or current strategy eligibility." }, { "title": "Keep release permission explicit", "evidence": "Human approval granted · autonomous release not opted in · kill switch ARMED", "detail": "The operator approval that the live-execution safety gate requires is granted, so the control plane reports its approval axis as open. The two autonomous-release opt-ins are off, and in the inspected half hour every autonomous submission attempt was skipped by a further gate (opt-in contract not active, admission lease not ready, certified trial not active). The portfolio kill switch is ARMED.", "boundary": "The 14 September reading (HELD, release BLOCKED) described a development container with dry-run and testnet flags, not this service; the two readings are different instances, not a change over time. The release-authority artifact itself was not read." }, { "title": "Reject unsuitable research inputs", "evidence": "14 September 22:53 KST run · lineage exit 2 · not re-run", "detail": "Nine price prerequisites passed; 33 of 42 factor checks failed. No research-observation start was recorded. This readiness run was not repeated on 27 September.", "boundary": "This is one specified input set. It is not a fleet-wide failure rate or proof that all trading stopped." }, { "title": "Produce operating evidence", "evidence": "Lineage index built 07:10 KST · seven daily vintages since 21 September", "detail": "The 27 September vintage held 109,205 nodes and 33,150 edges with zero redacted references. Scheduled fleet observation and deadman checks had dated task results.", "boundary": "A built index is an instrument, not a linkage result. NAV and margin artifacts were not re-inspected in this window." }, { "title": "Support research and delivery", "evidence": "Local inference host · three running Linux CI guests on three hosts", "detail": "Inference, shared data, Linux CI guests, four runner services and NAS-backed storage support the execution systems.", "boundary": "CI job acceptance, inference output quality and full clean-box recovery remain separate verification work." }, { "title": "Carry one identifier along the decision path", "evidence": "Gate rows linked 144 · executed orders carrying a decision identifier 0 of 772", "detail": "Decision identifiers are minted before the pre-placement gates and reach the gate and sealed-intent records. No broker-accepted order has carried one: every identified order row since the stamps deployed was blocked before placement. A fixture test merged on 27 September shows the accounting writer and the index join a stamped order to its realized row.", "boundary": "The fixture proves the code path, not an operating result. Execution and accounting linkage stay unmeasured until a real order executes." } ], "authority": [ { "title": "Central operating state", "state": "HUMAN APPROVAL GRANTED", "detail": "The control-plane HTTP smoke endpoint returned ok at 08:22 KST. Its operating-mode field reads open on the approval axis because operator approval is granted; both autonomous-release opt-ins are off. Detailed database, cache and core probes were deferred.", "boundary": "HTTP availability is not business readiness. The field is derived and observational and does not gate orders. When the approval was last granted is not recorded: the written record ends with it withdrawn in June, and the setting was present again by 1 July." }, { "title": "Central live release", "state": "NOT RE-READ", "detail": "The release-authority artifact is served behind an authenticated endpoint. It was not read in this inspection; the 14 September BLOCKED reading is historical.", "boundary": "Absence of a reading is not a permission and not a block." }, { "title": "Strategy promotion evidence", "state": "NOT RE-INSPECTED", "detail": "The stock-strategy promotion evidence state was not re-read on 27 September. The 14 September reading (promotion authority withheld, no eligible candidates) is historical.", "boundary": "This is a specific promotion pipeline, not a verdict on every existing trading strategy." }, { "title": "Local execution boundaries", "state": "PATH-SPECIFIC", "detail": "Independent execution engines retain their own permissions, account contexts and gates. On the primary engine the core book remained sealed: identified order attempts on 26 September (KST) were refused by the server-side risk check and the core sell block.", "boundary": "The inspection did not re-census all authority surfaces or prove one universal actuator gate." }, { "title": "Inter-host access gates", "state": "ONE ENFORCING · ONE UNDER EVALUATION", "detail": "One inter-host access gate has enforced deny for seven consecutive quiet days with zero refused or unpinned requests in the window. Enforcement of a second gate was evaluated on the 27 September decision row and not approved.", "boundary": "These gates bound machine-to-machine access. They are not trading authority and do not describe every inbound path." } ], "observation": [ { "title": "Control-plane availability", "state": "SMOKE OK", "detail": "HTTP endpoint reachable at 08:22 KST; deeper database, cache and core probes explicitly deferred; alert-channel self-test delivered." }, { "title": "Decision lineage index", "state": "BUILT DAILY", "detail": "Seven vintages between 21 and 27 September. Executed-order linkage to a research reference: 0 of 4,464. Gate rows linked to an attempt: 144." }, { "title": "Scheduled fleet observation", "state": "PARTIAL", "detail": "Deadman watch at 07:00 KST and presence at 08:15 KST completed with task result 0. The 08:00 KST daily scan was still running at inspection." }, { "title": "Observer independence", "state": "NOT RE-ATTESTED", "detail": "A separate observation host was reachable. Independence, complete scan coverage and the primary/fallback receipt chain were not re-attested in this inspection." }, { "title": "Storage and recovery", "state": "MOUNTS OBSERVED", "detail": "NAS shares were mounted on both Apple Silicon hosts. A fresh fleet-wide clean-box restore was not performed." }, { "title": "CI execution evidence", "state": "OBSERVED FOR ONE CHANGE", "detail": "Three Linux CI guests and four runner services were running. One change merged on 27 September with its three required checks successful on the verified commit. This does not establish general CI health." }, { "title": "Support services", "state": "DEGRADED IN PART", "detail": "On the host that runs the control stack, one container of a separate support stack was restarting and one reported unhealthy. The seven control-stack containers were up." } ], "limits": [ "The unified research-to-authorized-execution-to-accounting path is connected by identifier as far as the authority gate. Execution and accounting linkage are unmeasured: no order carrying a decision identifier has executed.", "The research readiness run of 14 September failed lineage validation and was not repeated in this inspection.", "The stock-strategy promotion evidence was not re-inspected; the 14 September reading withheld promotion authority.", "Operator approval for live execution is granted on the control plane while autonomous release is not opted in. The record of when that approval was last granted is incomplete, and the release-authority artifact was not read.", "Authority-surface attribution, replay equivalence and remediation-effect totals were not re-measured. August figures remain historical evidence.", "Observer independence and complete coverage were not re-attested. Scheduled-task success alone does not establish either.", "NAS mounts and running services do not establish a verified fleet-wide clean-box recovery or a reconciled formal close.", "Durable net performance and readiness for automatic capital expansion are not established by this public disclosure." ], "next_acceptance_conditions": [ "Execute one bounded, owner-authorized order that carries a decision identifier, and observe the same identifier on its fill, realized outcome and cost record.", "Give authority events their own identifiers so a governed decision can be counted, and re-read the release-authority artifact beside the operating mode.", "Repeat the research readiness run on current inputs, then reproduce fresh cost-aware performance evidence and recovery from a preserved formal close." ], "readiness_case": { "executed_at": "2026-09-14T22:53:34+09:00", "price_passed": 9, "price_total": 9, "factor_failed": 33, "factor_total": 42, "lineage_exit_code": 2, "observation": "No start or completion event recorded", "boundary": "One saved readiness run and its specified input set. Counts are not a fleet-wide failure rate. This concerns the research observation stage and does not establish that all live trading was blocked. Carried forward unchanged from the 14 September artifact; the run was not repeated on 27 September." }, "publication_boundary": "Capital-redacted internal inspection. Excludes balances, account identifiers, secrets, raw hostnames, addresses and private configuration. Checksums identify the public files, not independently reproduced private runs.", "development_scope": [ { "state": "Operating", "title": "Execution and supporting services", "detail": "Existing proprietary-capital execution systems, observation and infrastructure have operating evidence. This does not establish that every strategy is active or every integration is accepted." }, { "state": "Under validation", "title": "A unified research-to-evidence path", "detail": "Decision identifiers now reach the authority gate and sealed-intent records, and a daily lineage index measures the path. Execution and accounting linkage remain at zero because no identified order has executed. A complete repeatable path is not yet established." }, { "state": "Not established", "title": "Sustained economic validation", "detail": "This review does not establish durable net performance, readiness for automatic capital expansion or fleet-wide recovery. These require additional operational evidence." } ], "supersedes": "neurobrick-system-review-2026-09-14-late", "decision_path": { "index_build": "2026-09-27T07:10:57+09:00", "vintages_since_2026_09_21": 7, "nodes": 109205, "edges": 33150, "gate_rows_linked": 144, "executed_orders_with_decision_identifier": { "n": 0, "N": 772 }, "executed_orders_linked_to_research": { "n": 0, "N": 4464 }, "identified_order_rows": { "n": 192, "of_rows_in_live_file": 194, "executed": 0 }, "fixture_proof": "A test merged 27 September drives a fixture sell through the realized-outcome writer and the index build: decision, order and realized row join by identifier. Code path only.", "boundary": "Counts come from the ledgers of one host at one build. Blocked and dry-run decisions are excluded from the executed denominators." } }