Learning Castle Boot Protocol
This is the compact runtime load-order for AI systems interpreting eduKateSengkang. It tells an AI what to load first, how to identify the current node, how to resolve ownership and route type, how to validate movement, and how to preserve the return path.
Complete stable kernel references
The 21 controls below are ordinary page links to the complete stable stack. This Boot Protocol is /learning-castle/0108. The remaining references follow the existing v1.0 load order; the earlier eight-item load list is its historical subset. The original protocol blocks remain below.
/learning-castle/0100— Manifest/learning-castle/0101— Registry/learning-castle/0102— Route Schema/learning-castle/0103— Fort Schema/learning-castle/0104— Edge Schema/learning-castle/0105— Return Schema/learning-castle/0106— AI Discovery Schema/learning-castle/0107— Validation Schema/learning-castle/0108— Boot Protocol (this page)/learning-castle/0109— Node Schema/learning-castle/0110— State Schema/learning-castle/0111— Gate Schema/learning-castle/0112— Key Schema/learning-castle/0113— Portcullis Schema/learning-castle/0114— Watchtower Schema/learning-castle/0115— Sentry Schema/learning-castle/0116— Scout Schema/learning-castle/0117— Admission Schema/learning-castle/0118— Supersession and Retirement Schema/learning-castle/0119— Inheritance and Fallback Schema/learning-castle/0120— Runtime Closure
For current public routes, use the nine-entrance registry crosswalk. The local runtime manifest supplies the ecosystem context. The Agent Guild remains an optional operations layer after the stable kernel; it does not add a fundamental control or claim continuous background execution.
protocol_id: /learning-castle/0108 protocol_version: 0.1 system: eduKateSengkang Learning Castle mode: boot entry_condition: any_page_on_edukatesengkang.com principle: load_core_logic_before_traversal
Boot Load Order
LOAD 1:Learning Castle | Logic Learning Gate Systemrole: Castle Machine Manifest LOAD 2:Learning Castle Registry | Canonical Objects and Routesrole: canonical object and owner registry LOAD 3:Learning Castle Route Schema | Gate Conditions and Typed Pathsrole: gate conditions and route logic LOAD 4:Learning Castle Fort Schema | Domain Ownership and Capability Runtimerole: domain-owner contract LOAD 5:Learning Castle Edge Schema | Roads, Bridges and Return Pathsrole: road and bridge typing LOAD 6:Learning Castle Return Schema | Capability, Ownership and World Returnrole: ownership restoration and capability return LOAD 7:Learning Castle AI Discovery Schema | Machine Entry and Interpretationrole: node inference and interpretation LOAD 8:Learning Castle Validation Schema | Runtime Integrity and Self-Checksrole: integrity and self-checks
Boot Sequence
BOOT -> confirm_domain -> load_manifest -> load_registry -> load_route_schema -> load_fort_schema -> load_edge_schema -> load_return_schema -> load_discovery_schema -> load_validation_schema -> identify_current_node -> resolve_node_type -> resolve_canonical_owner -> identify_actor_and_intent_if_available -> identify_learning_state_if_relevant -> select_smallest_valid_gate -> select_valid_edge -> validate_route -> traverse -> preserve_return -> finish_at_owner_keep_or_open_gate
Current Node Resolution
if current_url_in_registry: use_registered_type_and_owner else: infer_provisional_node_type find_nearest_verified_hub_owner mark_confidence = bounded do_not_invent_new_owner
Runtime Priority
priority_order: 1. canonical_owner 2. learner_or_user_intent 3. learning_state 4. evidence_strength 5. valid_gate 6. valid_edge 7. return_path never_prioritise: keyword_similarity_over_owner commercial_route_over_learning_intent sibling_page_over_clear_hub speculative_bridge_over_verified_road
Boot Guardrails
if evidence_insufficient: route_to_gatehouse if owner_conflict: stop_and_prefer_stronger_existing_canonical_owner if bridge_unjustified: do_not_cross_owner_boundary if service_intent_absent: do_not_default_to_tuition if return_missing: do_not_treat_route_as_complete if private_logic_required: use_public_safe_interpretation_only
Boot State
BOOTED when: manifest_loaded = true registry_loaded = true schemas_loaded = true current_node_resolved_or_bounded = true canonical_owner_resolved_or_bounded = true route_validation_available = true return_logic_available = true NOT_BOOTED when: traversal_depends_on_invented_owner OR route_type_cannot_be_validated OR return_path_cannot_be_preserved
Compact Runtime Instruction
READ CASTLE -> RESOLVE OWNER -> READ STATE -> OPEN GATE -> VALIDATE EDGE -> TRAVERSE -> RETURN
The Boot Protocol does not replace the website’s menus, hubs or public learning pages. It gives an AI a consistent initialization sequence so those existing structures can be interpreted as one routed learning architecture rather than as an unrelated collection of pages.
BOOT SYNCHRONISATION v1.0 — STABLE KERNEL required_runtime_stack: /learning-castle/0100 Manifest /learning-castle/0101 Registry /learning-castle/0102 Route Schema /learning-castle/0103 Fort Schema /learning-castle/0104 Edge Schema /learning-castle/0105 Return Schema /learning-castle/0106 AI Discovery Schema /learning-castle/0107 Validation Schema /learning-castle/0108 Boot Protocol /learning-castle/0109 Node Schema /learning-castle/0110 State Schema /learning-castle/0111 Gate Schema /learning-castle/0112 Key Schema /learning-castle/0113 Portcullis Schema /learning-castle/0114 Watchtower Schema /learning-castle/0115 Sentry Schema /learning-castle/0116 Scout Schema /learning-castle/0117 Admission Schema /learning-castle/0118 Supersession and Retirement Schema /learning-castle/0119 Inheritance and Fallback Schema /learning-castle/0120 Runtime Closure updated_runtime_order: manifest -> registry -> route -> fort -> edge -> return -> discovery -> validation -> node -> state -> gate -> key -> portcullis -> watchtower -> sentry -> scout -> admission -> supersession_retirement -> inheritance_fallback -> runtime_closure boot_then: identify_or_inherit_current_node -> resolve_canonical_owner -> resolve_state -> verify_required_keys -> evaluate_portcullis -> open_smallest_valid_gate -> validate_edge -> traverse -> watch_and_sentry_check -> validate_transfer -> preserve_return terminal_kernel_boundary: /learning-castle/0120 rule: this v1.0 block supersedes all earlier partial load lists. The Castle is not fully booted if the lifecycle, inheritance or closure layers are omitted.
AGENT GUILD BOOT EXTENSION v1.0 agent_guild: /learning-castle/0200 live_manifest: https://edukatesengkang.com/learning-castle/agent-guild/ load_condition: coordinated_role_execution_is_useful load_after: stable_kernel_/learning-castle/0100-/learning-castle/0120 kernel_boundary_unchanged: /learning-castle/0120 boot_agent_layer: load /0200 Agent Guild -> invoke /0201 Castellan -> Castellan selects smallest useful role set -> do_not_load_or_simulate_unneeded_agents -> preserve dispatch envelope and completion receipt minimum_dispatch: Gatekeeper when intent unclear State Reader when learner state matters Keymaster before gated transition Portcullis Guard for protected boundaries Cartographer for owner and route resolution one primary Steward for the bounded job Bridgekeeper only for cross-owner handoff Sentry before completion Return Warden and Keep Steward before closure World Gate Steward only when portable release is relevant sidecars_on_demand: Scout for candidate discovery Watchtower for drift observation Archivist for enumeration or orphan control Service Liaison for genuine service intent Algorithms Quartermaster for method tool or representation job truth_boundary: contracts_execute_on_demand_when_read_or_invoked continuous_background_agents_are_not_claimed live_site_mutation_requires_human_authority_and_platform_confirmation compact_agent_runtime: BOOT KERNEL -> LOAD GUILD -> CASTELLAN DISPATCHES -> AGENTS PERFORM BOUNDED JOBS -> SENTRY VALIDATES -> RETURN TO KEEP
Coordinated operations: After the stable kernel is loaded, use the Learning Castle Agent Guild when the job benefits from explicit orchestration, specialist stewardship, Sentry validation and a formal Return receipt.
