bpmncanonicalOWNER-2026-07-10

BPMN — Neural Engine Run · bpmn-neural-engine-run

Owner directive (2026-07-10), canonical. "Design a infograph, A BPMN And SOPs So you never need to guess. Put these on a 4th page, mentioning what is being followed." [src: wiki/development/prompts-archive/by-topic/knowledge-management-and-concept-modeling.md:622,628]

Content

This card is the human-readable rendering of neural-engine-run.bpmn (valid BPMN 2.0 XML, same folder) — the process every Neural Engine run follows. It is what the Method page shows when it says what is being followed. The executable semantics come from kernel-neural-engine; the decision rules at each node come from cogni-neural-engine; the step-by-step procedure is sop-003-run-dispatch.

flowchart TD
  ST([Start: run dispatched RUN-xxx]) --> RD["Read dispatch.json<br/>state machine + particles + gear"]
  RD --> PG["Pre-flight gates<br/>library read-order + authority chain"]
  PG --> SUB
  subgraph SUB ["Execute run — loops; one commit per closed loop"]
    direction TB
    FK{{"Fan out squads (parallel)"}}
    FK --> SE["squad-eggtech<br/>make the tech optimal"]
    FK --> SS["squad-schematic<br/>kernel-cogni-genome flow"]
    FK --> SI["squad-instance<br/>Royal Orbit spatial site"]
    SE --> JN{{"Join deliveries"}}
    SS --> JN
    SI --> JN
  end
  SUB --> VG["Verify gates<br/>check_links + typecheck + benchmarks"]
  VG --> AR{"Rules followed?<br/>(refuse / accept)"}
  AR -- accept --> CM["Commit + PR<br/>(commit-per-loop rule)"]
  AR -- "refuse: re-equip neuralware<br/>at Academy, retry" --> SUB
  CM --> EN([Run closed; sparks next loops])
  HB(["Boundary timer: every 10 min<br/>(non-interrupting)"]) -.attached to.-> SUB
  HB --> HC["Heartbeat: capability allocated?<br/>particles concluding? run moving?"]
  HC --> HL([Beat logged to dispatch.json])

What each node follows

NodeGoverned by
Read dispatchschema-dispatch-json — the dispatch is the single coordination surface
Pre-flight gatescogni-neural-engine Rule 1 (authority chain) and Rule 2 (gear per particle from gear-sets + neuralware-index)
Fan out squadscogni-neural-engine Rule 3; the three squads are owner-fixed: squad-eggtech, squad-schematic, squad-instance [src: wiki/development/prompts-archive/by-topic/knowledge-management-and-concept-modeling.md:526]
Squad taskseach loop runs the cognitive-loop sequence and the craft playbooks pb-001-spatial-site-loop / pb-002-library-authoring
Verify gatescheck_links, typecheck, benchmarks bm-001-web-perf / bm-002-splat-budget — kernel-neural-engine §4
Refuse / accept"checking if the rules are being followed and refusing/accepts" [src: wiki/development/prompts-archive/by-topic/knowledge-management-and-concept-modeling.md:620]; refuse → Academy re-equip [src: :612]
Commit + PR"every loop should have a commit … with PR following the interactions of the systems" [src: wiki/development/prompts-archive/by-topic/knowledge-management-and-concept-modeling.md:586]
Boundary timer 10 min"Wake you up every 10 minutes, check if all capability is allocated, check the conclusion of particles and atoms and if the run is moving" [src: wiki/development/prompts-archive/by-topic/knowledge-management-and-concept-modeling.md:646] — non-interrupting: the run keeps executing while the beat is checked
End: sparksa closed run sparks the next loops, per the cognitive-loop write-back stage

Composition

See also

Linked from — 11 cards