EGGTECH Squad · squad-eggtech
Owner directive (2026-07-10), canonical. "Make 3 squads work, one with a run for the eggtech (making the tech optimal) another for the Schematic (…) and another one for instance (…)." [src: wiki/development/prompts-archive/by-topic/knowledge-management-and-concept-modeling.md:526]
Content
EGGTECH is one of the three dynamic squads the owner ordered stood up for the HOS-Instance build. Its mission, verbatim, is "making the tech optimal" [src: wiki/development/prompts-archive/by-topic/knowledge-management-and-concept-modeling.md:526]. Operationally that scope covers:
- Engine — the spark-renderer / gaussian-splatting runtime, the holonprint render loop, WebGL/Three integration, and the pad/pod conversion path ("maybe now convertions from PADs PODs with more PADs over and even efficient PODs, such as html5 or whatever is lighter" [src: wiki/development/prompts-archive/by-topic/knowledge-management-and-concept-modeling.md:459]).
- Performance — owning and enforcing the hard gates in bm-001-web-perf and bm-002-splat-budget, against the owner's mandate that "after the loading the experience is as fluid and light as scientifically possible" [src: …/knowledge-management-and-concept-modeling.md:459] on any device, "doesnt matter the device, 15 years old mobile" [src: …/knowledge-management-and-concept-modeling.md:463].
- Tooling — build pipeline (cb-vite), CI gates, QA hooks, dispatch tick instrumentation (sop-003-run-dispatch).
Its work is executed as RUN-017, dispatched and heartbeat-checked per sop-003-run-dispatch; every loop it runs follows sop-001-loop-execution and traverses the cognitive-loop.
Seats
The six knowledge-role seats follow the owner's role mapping (see squad-roles) [src: …/knowledge-management-and-concept-modeling.md:552-557]. Canonical datan names exist only for two seats: Assistant = "Jules" (D24) and Companion = "Modo" (D25) [src: HOS-Webdoc/web/src/data/hos-specification.ts:452, 459]; remaining seats are named at run-open.
| Seat | Mobilizer | Datan | Default gear-set (gear-sets) |
|---|---|---|---|
| Knowledge Engineer | dooit | assign at run-open | splat-authoring, build-tooling |
| Knowledge Analyst | ditto | assign at run-open | build-tooling (RAG/gateway/cache duty [src: …/knowledge-management-and-concept-modeling.md:561]) |
| Knowledge Management | assistant | Jules (D24) | run-orchestration |
| Knowledge Architect | companion | Modo (D25) | run-orchestration |
| Knowledge Executive | advisor | assign at run-open | run-orchestration (benchmark verdicts) |
| Knowledge Designer | construct | assign at run-open | scroll-motion |
Run
- Run id: RUN-017 (a Run is "a one-month delivery unit aggregating cycles, nodes and particles into a single deliverable the success manager signs off" [src: wiki/GLOSSARY.md:2707]).
- Deliverable: all bm-001-web-perf + bm-002-splat-budget gates green on the Royal Orbit spatial-site; tooling stable for the other two squads.
- Failure handling: a datan that cannot deliver swaps neuralware per the academy rule in sop-001-loop-execution.
Composition
- uses: gear-sets (deterministic loadout per seat) · bm-001-web-perf / bm-002-splat-budget (owned gates) · pb-001-spatial-site-loop (playbook rows it executes) · cb-spark / cb-threejs / cb-vite (cookbooks) · sop-001-loop-execution / sop-003-run-dispatch (procedures)
- used-by: okrs-2026-h2 (Initiative I-1) · squad-instance (consumes its engine + budgets) · squad-schematic (consumes its tooling)
See also
- squad-roles — the seat/role mapping shared by all three squads
- squad-schematic · squad-instance — sibling squads (RUN-016, RUN-018)
- neural-engine — the runtime the squad's loops execute on