NODEDC_1C/docs/ARCH/11 - architecture_turnaround/11 - continuity_stabilization_plan_2026-04-17.md

11 - Continuity Stabilization Plan (2026-04-17)

Purpose

This note defines the recovery plan for the current pre-expansion breakpoint.

The goal is not to patch individual failing prompts.

The goal is to finish the missing runtime authority that should govern mixed live sessions after the turnaround 11 owner extractions.

Current Reading

The strongest current evidence is:

• narrow and company-selected scenarios can pass end-to-end;
• mixed saved-session runtime still fails on root inventory, selected-object continuity, same-date restore, and cross-domain same-date pivot;
• therefore the architecture is not missing only routes;
• it is missing one governing continuity authority.

In one sentence:

• decision ownership became distributed faster than continuity ownership became explicit.

What This Plan Stabilizes

This plan is specifically about one system object:

• assistant_session_continuity_v1

That object should become the shared authority for:

• active root frame
• active selected object
• active organization scope
• active date scope
• active clarification state
• active answer object / reusable bundle
• recap source of truth

Target Runtime Rule

Before any of the following decisions are made:

• route arbitration
• company clarification
• selected-object follow-up routing
• same-date restore
• recap answer generation

the runtime must first resolve one continuity snapshot for the active session.

Those downstream owners may interpret the snapshot differently, but they must not reconstruct competing versions of the session state independently.

Immediate Passes

Pass A. Install shared continuity snapshot

Scope:

• create one shared continuity resolver for session items and grounded address context;
• centralize extraction of active item, organization, date, root frame, and recap-eligible grounded context;
• stop allowing recap and adjacent meta logic to build over ungrounded or clarification-only state.

Exit condition:

• recap can only trigger over verified grounded address context;
• selected-object memory cannot be reconstructed from failed clarification turns;
• route and memory layers consume the same continuity snapshot.

Pass B. Reduce clarification priority conflicts

Scope:

• move clarification behind restored continuity when the business frame is already sufficient;
• stop repeated company clarification from interrupting same-family continuation;
• make clarification state explicit and resumable instead of re-discovered ad hoc.

Exit condition:

• repeated clarification no longer appears after adjacent grounded business answers in the same thread;
• selected-object and same-date follow-ups stop falling into generic company templates.

Pass C. Re-ground recap and answer packaging

Scope:

• recap must summarize verified session facts only;
• answer packaging must not sound more certain than truth assembly;
• technical scaffolding must not leak into the top block of user-facing answers.

Exit condition:

• recap cannot claim supplier/date/document facts that were never grounded;
• meta boundary replies no longer expose MCP, read-only, route ids, capability ids, or debug labels;
• user-facing top blocks remain business-first.

Pass D. Lock mixed runtime as the primary gate

Scope:

• promote mixed saved-session runtime to the main architecture gate before domain expansion;
• keep narrow harnesses and seam tests, but do not let them overrule mixed replay;
• evaluate critical user paths rather than isolated route green status.

Exit condition:

• the core mixed replay is green on direct answer, selected-object continuity, same-date carryover, recap truthfulness, and technical cleanliness;
• no unresolved P0 remains on the primary user path.

Anti-Goals

This stabilization pass is not:

• a rollback to the old monolith
• a case-by-case regex patch sweep
• a prompt-only wording cleanup
• a UI-only improvement pass

Practical Sequence

1. Finish the continuity snapshot and wire it into the hot route / recap path.
2. Rework clarification precedence so it becomes a last meaningful step.
3. Harden recap and boundary presentation against ungrounded and technical output.
4. Rerun the mixed AGENT replay until the critical continuity edges are green.
5. Only then continue deeper intent extraction and wider domain expansion.

Current Pass Status

Completed in the current working pass:

• shared continuity snapshot is already wired into recap and adjacent route memory logic;
• grounded address history can now restore active organization scope instead of depending only on explicit company-selection metadata;
• early organization clarification no longer outranks item-focused inventory follow-up paths when the session already carries a strong object frame;
• meta boundary replies were already cleaned from technical MCP / read-only leakage.
• early non_domain arbitration no longer suppresses a positive L0 address-lane decision for colloquial but supported exact routes;
• foreign-accounting pivots over inventory drilldown now preserve root-scoped carryover instead of dropping continuity before root-frame sanitation;
• the wide assistantAddressFollowupContext regression pack is green again, including month-only VAT follow-up and inventory -> VAT pivot sanitation.
• counterparty document root wording is now recovered through unicode-safe exact signals instead of depending on mojibake-sensitive legacy phrases;
• declined Russian account wording like по счёту 60 now restores account scope inside polarity/runtime guards instead of collapsing into other_numeric;
• exact address intents can now stay in the address lane even if the semantic guard overflags deep investigation without an actual investigative user request;
• selected-object inventory follow-ups can now override a stale stock root intent when the semantic contract already marks selected_object_scope_detected, including exact user wording like по выбранному объекту ... где взяли это;
• explicit capability-meta wording for дельта по договорам now keeps the asked capability in the user-facing answer instead of collapsing into the generic что ты умеешь catalog reply.
• the transition hot path now starts consuming the shared continuity snapshot as fallback authority for active item / active organization / grounded inventory root frame instead of rebuilding those values only from local ad hoc history scans;
• live replay address_truth_harness_phase7_meta_domain_mix_live_20260417_post_arch_fix_rerun2 is accepted end-to-end with 14/14 steps green, including the previously broken step_01_counterparty_documents and step_04_open_items_account_60.

Still open after this pass:

• mixed continuity is now strong enough for the current phase7 gate, but it still needs broader saved-session proof before domain expansion can be treated as low-risk;
• the next architecture pass should move from one repaired mixed replay to a wider saved-session set and multi-domain acceptance pack;
• remaining work should focus on keeping the unified continuity authority stable under new real user paths, not on wording-only polish or isolated route greens.
• company authority is still not proactive enough at root inventory entry in multi-company sessions without an already grounded active organization;
• the next stabilization slice should prefer system-level company authority handling over repeated local clarification templates when the session has enough business context.

Completed in the current follow-up pass:

• direct company activity-age wording like а по Альтернативе Плюс сколько лет активности в базе 1С? is now protected by a unicode-safe exact signal instead of depending on mojibake-sensitive legacy lifecycle phrases;
• capability meta answers now explain supported business groups through human examples instead of leaking internal operation ids like vat_period_snapshot, inventory_on_hand_as_of_date, explain_boundary, or suggest_safe_next_step;
• the next proof target after unit/build checks is the live phase5 replay, because it exercises both the restored activity-age path and the capability-meta interrupt in one shared session.

Latest live replay evidence after that proof run:

• the capability meta interrupt is now business-first and no longer leaks internal operation ids in the top block;
• the same replay exposed a stricter continuity defect that the top-level review initially missed: organization identity can drift in session state as a damaged live label like ООО \\Альтернати"а Плюс\\;
• when that happens, the runtime keeps both organization and a stale counterparty anchor, does not emit counterparty_cleared_for_selected_organization_activity, and falls into counterparty_anchor_not_matched_in_materialized_rows;
• this is a system-level organization-identity robustness gap between data-scope probing, continuity memory, and exact-route truth gating, not a wording-only prompt defect;
• the current stabilization slice therefore includes hardening organization identity matching itself and rerunning the same live pack until step-level human answers and review verdicts align.

Latest phase8 runtime authority evidence after the manual mixed replay hardening:

• live replay address_truth_harness_phase8_manual_runtime_authority_mix_live_20260417_rerun1 proved that the activity-age route was restored, but also exposed a hidden false-green: step_11_inventory_same_date_after_receivables silently reused stale inventory-root date 2021-03-31 instead of the freshest receivables date 2020-03-31;
• the first fix in assistantService was not sufficient on its own, because decomposeStage still rebuilt inventory_root follow-up context by overwriting previous_filters from root_filters wholesale;
• the architectural correction was to preserve root authority for organization / warehouse while preserving the freshest temporal scope (as_of_date, period_from, period_to) from the immediately previous grounded step;
• this was locked by direct regressions in assistantTransitionPolicy.test.ts and addressInventoryRootFrameRegression.test.ts, plus a live rerun against the same manual replay spec;
• live replay address_truth_harness_phase8_manual_runtime_authority_mix_live_20260417_rerun4 is now accepted end-to-end with 14/14 steps green, including:
  • step_07_capability_meta with business-first human wording;
  • step_11_inventory_same_date_after_receivables on the correct date 31.03.2020;
  • step_14_company_activity_age with restored factual lifecycle answer;
  • cleaned user-facing company labels in the data-scope meta reply (ООО Альтернатива Плюс, ООО Лайсвуд, РАЙМ) instead of damaged raw probe labels.

Still open after the accepted phase8 replay:

• proactive organization authority at the very beginning of a new multi-company bookkeeping session is still weaker than the target product feel; the current system now clarifies honestly and cleanly, but it does not yet always pre-offer company selection early in the conversational flow;
• some user-facing inventory/counterparty labels inside business answers still deserve final presentation cleanup, but these are now post-stabilization quality refinements rather than continuity-authority blockers.

Latest phase9 proactive-authority evidence after the fresh multi-company replay:

• a new live replay address_truth_harness_phase9_proactive_scope_offer_live_20260418_rerun3 is accepted end-to-end with 5/5 steps green;
• on the very first smalltalk turn, the assistant now stays in normal living-chat mode but appends a business-first proactive organization offer instead of waiting for a later forced clarification;
• explicit company choice in the next turn is now fixed deterministically into session authority before the first accounting route, so later business turns inherit one stable active organization;
• the restored activity-age route for ООО Альтернатива Плюс is now proven again inside the same shared session, not only in isolated route checks;
• the previously broken same-date inventory pivot after receivables is now routed as inventory_on_hand_as_of_date with the carried date 31.03.2020 and the carried organization ООО Альтернатива Плюс, without falling back into repeated company clarification;
• this phase therefore closes the remaining gap called out at the end of phase8: proactive company authority is no longer purely reactive in fresh multi-company bookkeeping sessions.

Still open after the accepted phase9 replay:

• business answers are now semantically correct on this path, but some inventory list formatting still feels heavier and more mechanical than the target human style;
• the next architecture slice should keep expanding saved-session proof across additional real user chains, while separately tightening answer presentation so exact routes do not feel template-driven even when the truth path is already correct.

Latest phase10 bridge-and-aggregate evidence after the manual replay recovery:

• live replay address_truth_harness_phase10_manual_bridge_and_aggregate_mix_live_20260418_rerun8 is accepted end-to-end with 9/9 steps green;
• the previously broken bridge selected item purchase provenance -> VAT on purchase date is now explicit instead of implicit:
  • the continuity layer derives the purchase month from the grounded provenance evidence;
  • the same session keeps selected object continuity instead of collapsing into generic root-only VAT arbitration;
  • the runtime now routes this follow-up as vat_liability_confirmed_for_tax_period, not as forecast, unknown, or generic clarification;
• the same replay also proves that the neighboring aggregate fixes are live on the real assistant path:
  • top-customer-all-time now returns a direct business answer first;
  • top-year aggregate now returns a direct business answer first;
  • very-old-stock now prefers inventory_aging_by_purchase_date over a generic inventory snapshot;
• this matters architecturally because the seam that used to exist only as ambient monolith behavior is now protected as an explicit carryover contract plus replay-backed acceptance path.

Still open after the accepted phase10 replay:

• the user-facing VAT explanation block is now correct and grounded, but some long exact answers still feel heavier than the target human product tone;
• the next architecture slice should keep moving from repaired bridge authority into answer-shaping cleanup and broader saved-session replay coverage, not back into isolated wording tweaks.

Latest phase11 manual follow-up/meta-quality evidence after the current hardening loop:

• live replay address_truth_harness_phase11_manual_followup_meta_quality_live_20260418_rerun6 is accepted end-to-end with 10/10 steps green;
• the previously broken ты умеешь считать дельту по договорам? branch is now protected by an explicit authority rule:
  • raw capability-meta intent outranks canonical predecompose rewrites that look like address retrieval;
  • stale VAT follow-up continuity no longer wins over a fresh capability/meta question in the same session;
• the previously broken short counterparty retarget а по свк is now clean on the real assistant path:
  • the display label uses the most specific confirmed counterparty name instead of a generic group fallback or a stale carryover anchor;
  • short uppercase Cyrillic acronyms like СВК no longer get stripped by the user-facing sanitizer as false mojibake;
  • the replay acceptance rule now targets the real regression (Контрагент: Группа Найдено ...) instead of incorrectly rejecting valid names like Контрагент: Группа СВК.;
• this phase matters architecturally because it closes two different seam classes at once:
  • meta authority vs stale follow-up authority;
  • resolved business label vs boundary sanitization noise.

Still open after the accepted phase11 replay:

• the current phase11 path is now semantically clean, but broader manual/user session packs still need to be replayed before expansion can be called low-risk across new domains;
• answer shaping on some long exact list answers is still heavier than the target human product feel, even though the truth path and routing are now correct;
• the next architecture slice should move to wider saved-session acceptance coverage and humanized exact-answer presentation, not back to isolated prompt-level repairs.

Latest continuity-authority convergence evidence after the current route pass:

• the route hot path now consumes the shared continuity snapshot directly instead of relying only on local findLastGrounded... helpers:
  • grounded address context can now survive into route arbitration even when the legacy local helper returns nothing for the current turn shape;
  • active organization continuity is now allowed to participate in organization-selection arbitration, instead of forcing route policy to reconstruct that context only from immediate clarification payloads;
• a bare organization-selection turn after grounded bookkeeping continuity is no longer automatically classified as non_domain_query_indexed noise when the session still carries valid grounded business context;
• session organization recovery inside the data-scope layer now has a final fallback to the same continuity snapshot, reducing one more duplicate path that used to rescan assistant history independently;
• the living-chat runtime now also consumes continuity-backed organization authority:
  • deterministic organization-fact boundary replies can now trigger from grounded continuity even when sessionScope.selectedOrganization and sessionScope.activeOrganization are both empty at runtime entry;
  • the chat layer now records whether it entered with grounded continuity and which organization came from that continuity snapshot, making future saved-session review less blind;
  • proactive organization offer logic is now explicitly blocked when grounded address continuity already exists, so the chat layer does not re-offer company selection on top of an already grounded business session;
• the next continuity-authority pass now centralizes organization authority in the shared continuity policy instead of letting route, living-chat, and data-scope reconstruct it independently:
  • a single resolver now merges grounded address continuity, assistant-side living-chat scope memory, session scope, and clarification candidates into one authority object;
  • route arbitration now consumes that shared authority for organization clarification continuation instead of rebuilding candidates and fallback selection locally;
  • living-chat runtime now boots with the same authority object, so knownOrganizations, selectedOrganization, and activeOrganization no longer start from a different merge order than the route layer;
  • data-scope active-organization recovery now also reads the same authority object, reducing one more history-scan seam that used to drift from route and chat;
  • targeted tests now protect this precedence layer directly, including the case where assistant-side living-chat selection must outrank older grounded continuity without losing the grounded organization in the merged known set;
• the first human-answer-shaping cleanup pass is now applied to heavy profile/aggregate exact answers:
  • period_coverage_profile and document_type_and_account_section_profile now start with a direct business-first lead (Коротко: ...) instead of service-flavored intros like профиль собран / строк агрегата;
  • the top block now states the business conclusion first and leaves ranked detail blocks below, which reduces the catalog-like feel without hiding the actual data;
• the next human-answer-shaping cleanup pass is now applied to VAT exact replies:
  • vat_payable_forecast and vat_liability_confirmed_for_tax_period now open with a business-first Коротко: ... lead, while the detailed calculation stays in the secondary block;
  • service-flavored top lines like Собран прогноз..., Режим результата..., and Строк агрегата... are removed from the first screen of the reply, which makes VAT answers read like user-facing guidance instead of an engine report;
  • VAT reply tests now explicitly protect this top-block shape, so future changes cannot silently reintroduce the same mechanical preamble;
• the next human-answer-shaping cleanup pass is now applied to counterparty ranking/profile replies:
  • counterparty_activity_lifecycle, contract_usage_overview, customer_revenue_and_payments, supplier_payouts_profile, and contract_usage_and_value now open with business-first wording instead of service-flavored профиль собран / строк агрегата / строк источника;
  • ranking and contract replies now preserve user wording better in the visible heading layer, including минимальный бюджет phrasing for low-turnover active contracts;
  • targeted ranking/profile tests now protect the new top-block shape, so these families are less likely to regress back into report-like wording during later route/domain work;
• the next human-answer-shaping cleanup pass is now applied to plain list replies in the exact lane:
  • list_contracts_by_counterparty, list_documents_by_contract, bank_operations_by_counterparty, bank_operations_by_contract, and the generic factual-list fallback no longer leak live address lane / catalog address lane wording into the user-facing answer;
  • these list replies now start with direct business-first leads and keep the selected rows below, which preserves factual usefulness without exposing internal routing labels;
  • targeted utf8 header tests now explicitly protect against lane leakage in these list families;
• the next human-answer-shaping cleanup pass is now applied to open-contract evidence-heavy replies:
  • list_open_contracts and open_contracts_confirmed_as_of_date no longer open with numbered Блок 1/2/3... report framing and now start with direct business-first summaries;
  • section headings are still structured, but they now read like user-facing guidance instead of an internal audit report, while keeping the same factual slices and evidence detail below;
  • targeted open-contract tests now protect the no-Блок 1 top-block shape, so future contour work cannot silently bring the report framing back;
• the next human-answer-shaping cleanup pass is now applied to confirmed debt snapshots:
  • payables_confirmed_as_of_date and receivables_confirmed_as_of_date now open with business-first Коротко: ... summaries instead of numbered report framing;
  • debt snapshot sections now keep the same factual structure, but top-level headings are user-facing (Что учтено, Сводка, Категории...) rather than Блок 1/2/3...;
  • direct compose tests now protect the no-Блок 1 top-block shape for both confirmed debt families;
  • isolated runtime proof for the payables_confirmed_as_of_date tryHandle path still needs a wider rerun, because the narrow harness invocation currently returns undefined before semantic assertions and therefore is not reliable evidence for this shaping pass by itself;
  • this is still not the end of shaping work: heuristic debt shortlists and some residual catalog-style blocks still need the same cleanup;
• this pass does not yet finish full single-owner continuity, but it narrows one of the remaining seams where route arbitration and scope memory could disagree about whether the session was still grounded.

Latest phase12 wider saved-session replay evidence after the current architecture pass:

• a new live replay address_truth_harness_phase12_wider_saved_session_pool_live_20260418_rerun3 is accepted end-to-end with 20/20 steps green;
• this wider pack now proves one shared assistant session across a longer mixed path:
  • first-turn smalltalk with proactive company offer;
  • explicit company fixation;
  • capability-meta interruption;
  • inventory roots and historical roots;
  • selected-object provenance and sale follow-up;
  • selected-item -> VAT-period bridge;
  • revenue aggregate pivot;
  • payables/receivables polarity mirror;
  • counterparty documents and short-name follow-up;
  • account-60 tails;
  • inventory aging and company activity-age;
• the replay exposed one real remaining seam before final acceptance:
  • colloquial smalltalk entry reached the correct living_chat + proactive_scope_offer lane, but the first answer still depended on raw LLM preamble and leaked irrelevant generic chat (Какой сегодня день?);
  • the fix was to convert first-turn smalltalk + proactive organization offer into a deterministic guarded entry path instead of trusting uncontrolled LLM preamble above the offer;
  • colloquial living-chat detection is now unicode-safe and no longer depends on Cyrillic \\b boundaries that silently fail in JavaScript regex;
• this matters architecturally because another formerly ambient monolith behavior is now an explicit runtime contract:
  • first-turn proactive company entry is controlled by runtime authority, not prompt luck;
  • the broader saved-session pack now proves that organization authority, continuity carryover, and top-level chat entry survive one longer real-user trajectory rather than only the earlier flagship chains.

Still open after the accepted phase12 replay:

• replay breadth is now better than before, but still not yet broad enough to call multi-domain expansion low-risk by default;

• the biggest remaining architecture risk is no longer the original continuity collapse, but the unfinished convergence toward one true single-owner session authority across every hot path;

• the next execution slice should therefore continue prioritizing:

  • more saved-session and multi-trajectory replay breadth;
  • less duplicated state reconstruction in route / transition / living-chat glue;
  • controlled expansion only after those broader proofs stay green.

• the next authority-convergence pass now also removes one more duplicate scope builder at the very start of the turn:

  • assistantOrganizationScopeRuntimeAdapter no longer reconstructs knownOrganizations / selectedOrganization / activeOrganization only from its own local history helpers;
  • the same session-scope bootstrap now consumes the shared continuity-backed organization authority first and only then falls back to older helper outputs and navigation-state overrides;
  • this matters because route, living-chat, data-scope, and early session-scope bootstrap are now closer to the same organization merge order instead of keeping a separate pre-route branch with its own drift risk;
  • targeted tests now explicitly protect the case where assistant-side continuity authority is present in prior assistant debug even when legacy history helpers are empty;
  • live replay address_truth_harness_phase12_wider_saved_session_pool_live_20260418_rerun4 remains accepted 20/20, which is the critical proof that this authority convergence did not reopen the wider saved-session path.

• the next architecture pass exposed and fixed a deeper hybrid-investigation contract break outside the flagship address chain:

  • phase13 replay showed that hybrid anomaly/tails questions were no longer failing only on follow-up continuity; the root entry itself had stopped entering the hybrid lane because deep-turn normalization was calling the normalizer with promptVersion=address_query_runtime_v1;
  • the model was already returning valid normalized_query_v2_0_2 payloads for these questions, but the runtime validated them as legacy v1, discarded the parsed normalized object, and then dropped into claim-bound clarification with empty fragments / route_summary;
  • deep-turn normalization now forces an actual normalizer contract (promptVersion=normalizer_v2_0_2, schemaVersion=v2_0_2) unless the caller already supplied a normalizer-family prompt explicitly;
  • this is a real architecture fix, not a case patch: deep-turn no longer inherits a user-facing address runtime prompt as if it were a normalizer schema contract;
  • targeted deep-turn normalization tests now protect this seam directly, including the case where an address-lane prompt would previously have caused silent v1 validation and total fragment loss;
  • live replay address_truth_harness_phase13_hybrid_followup_authority_live_20260418_rerun3 is now accepted:
    • root supplier tails anomaly questions re-enter hybrid_store_plus_live with grounded fragments and non-empty deterministic route summaries;
    • narrowing follow-up for 2020-06 / account 60 now keeps hybrid/batch routing instead of collapsing into empty clarification;
    • the broader hybrid investigation contour is therefore back under explicit runtime authority rather than ambient luck.

• the remaining translit root seam is now also closed in the same contour:

  • transliterated supplier-tail wording no longer loses the causal tail during predecompose entry handling;
  • live replay address_truth_harness_phase13_hybrid_followup_authority_live_20260418_rerun4 is accepted with the translit root step returning factual_with_explanation and staying inside hybrid investigation routing;
  • endpoint coverage now explicitly requires the translit account-60 tail question to keep every routed fragment in hybrid_store_plus_live, so future refactors cannot silently split the same question back into hybrid + store_canonical.

• the next authority-convergence pass now removes one more local organization reconstruction seam from the transition hot path:

  • assistantTransitionPolicy no longer reconstructs clarification/company authority only from ad hoc history scans and raw continuity snapshot pieces;
  • follow-up carryover now reads the shared organization authority object first, including assistant-side active organization memory and clarification candidates, before falling back to older local filters;
  • this matters because mixed follow-up questions that pivot after assistant-side company fixation no longer depend on whether the previous address debug happened to still carry organization in its own extracted filters;
  • targeted transition regression now protects the case where grounded history is empty but assistant-side organization authority is already present;
  • wide saved-session replay address_truth_harness_phase12_wider_saved_session_pool_live_20260418_rerun5 remains accepted 20/20, which is the critical proof that this transition-layer convergence did not reopen the broader continuity path.

• the next active-context convergence pass now removes one more duplicate address debug -> item/date/organization parser from memory-recap and living-chat follow-up builders:

  • assistantContinuityPolicy now exposes one shared helper for item, organization, and scopedDate extraction from grounded address debug, including root-frame fallback;
  • assistantMemoryRecapPolicy now consumes this shared helper instead of manually rebuilding the same context from extracted_filters and address_root_frame_context in multiple places;
  • this matters because deterministic memory-recap and historical-inventory capability replies now depend on the same context interpretation as the rest of continuity policy, rather than on a separate local parser that could drift on root-frame-only turns;
  • targeted continuity / memory-recap / living-chat tests now protect the root-frame fallback path explicitly;
  • wide saved-session replay address_truth_harness_phase12_wider_saved_session_pool_live_20260418_rerun6 remains accepted 20/20, which is the critical proof that this context-helper convergence did not reopen the broader living-chat continuity path.

• the next cleanup pass also removes one more class of false owners from the living-chat adapter itself:

  • assistantLivingChatRuntimeAdapter no longer keeps local dead history scanners for grounded inventory / selected-object / generic address debug lookup that are not part of the active execution path anymore;
  • this does not change runtime behavior directly, but it reduces the chance that future fixes accidentally revive or patch a stale local owner instead of the shared continuity / memory-recap policy seam;
  • targeted living-chat adapter tests and backend build remain green after the cleanup, which is the necessary proof that this was a structural owner-reduction pass rather than a hidden behavior change.

• the next continuity-authority pass now removes one more local addressDebug -> carryover anchor/date parser from the transition hot path:

  • assistantContinuityPolicy now exposes shared helpers for anchorType/anchorValue resolution and raw temporal carryover scope (as_of_date / period_from / period_to) from grounded addressDebug, including root-frame fallback;
  • assistantTransitionPolicy now consumes these helpers instead of rebuilding previous anchor selection from raw anchor_value_* and filter fields inline, and instead of reading carryover dates directly from readAddressDebugFilters(...) in multiple ad hoc places;
  • this matters because follow-up carryover is now closer to the same continuity interpretation layer that already owns item / organization / scoped-date facts, rather than keeping a separate transition-local parser for the same runtime evidence;
  • targeted continuity and transition regressions now protect inferred anchor carryover when explicit anchor_type is absent, plus root-frame temporal fallback at the helper layer;
  • wide saved-session replay address_truth_harness_phase12_wider_saved_session_pool_live_20260418_rerun7 remains accepted 20/20, which is the critical proof that this transition-layer convergence did not reopen the broader saved-session path.

• the next organization-authority pass now removes one more local history parser from the data-scope / session-bootstrap contour:

  • assistantDataScopePolicy.extractKnownOrganizationsFromHistory no longer rebuilds known organizations from raw assistant_* / organization_candidates / extracted_filters / root_frame_context fields on its own;
  • it now reads the shared resolveAssistantOrganizationAuthority(...) result first and only keeps assistant-text parsing as a compatibility fallback for older free-text scope replies;
  • this matters because data-scope probing, organization selection bootstrap, and the broader continuity layer are now closer to one canonical organization merge order instead of keeping a separate debug-field collector inside the scope policy;
  • targeted data-scope and organization-scope tests now protect that known organizations still include assistant-side authority, grounded address context, and free-text fallback organizations;
  • wide saved-session replay address_truth_harness_phase12_wider_saved_session_pool_live_20260418_rerun8 remains accepted 20/20, which is the critical proof that this data-scope convergence did not reopen the flagship saved-session path.

• the next session-bootstrap pass now removes one more legacy callback seam from the organization scope adapter:

  • assistantOrganizationScopeRuntimeAdapter no longer depends on a separate findLastAssistantActiveOrganization(...) callback even though it already computes shared continuity-backed organization authority internally;
  • active organization bootstrap now flows through selected organization, navigation organization, and shared continuity authority in one place instead of keeping a second callback-shaped fallback branch beside the authority object;
  • assistantService.resolveSessionOrganizationScopeContext(...) no longer passes that legacy callback into the runtime adapter, which reduces one more orchestration seam where old and new organization owners could drift;
  • targeted organization-scope, data-scope, and route regressions remain green after the change, and wide saved-session replay address_truth_harness_phase12_wider_saved_session_pool_live_20260418_rerun9 remains accepted 20/20, which is the critical proof that this bootstrap convergence did not reopen the flagship continuity path.

• the next replay-breadth pass now proves a different late-session contour instead of replaying only the flagship chain:

  • a new live pack address_truth_harness_phase14_counterparty_tail_resume validates data-scope meta -> explicit company selection -> counterparty docs -> short-name follow-up -> inventory today -> account 60 -> inventory aging -> historical inventory -> organization activity analytics inside one shared session;
  • the first draft of this pack exposed one real architecture seam rather than another continuity collapse: Как ты оценишь деятельность компании? after grounded organization activity-age was still falling into living_chat because the route depended too much on the L0 gate and not enough on the resolved supported intent;
  • addressCounterpartyIntentSignals now treats company-activity assessment wording as the same counterparty_activity_lifecycle contour instead of leaving it as unsupported meta chat;
  • assistantRoutePolicy now recovers the address lane from a supported resolved intent even when the initial L0 gate stays negative, so the system no longer loses a real business contour just because the low-level shape classifier stayed unsupported;
  • targeted counterparty UTF-8 and route-policy regressions now explicitly protect this seam, including the exact late-tail wording Как ты оценишь деятельность компании?;
  • live replay address_truth_harness_phase14_counterparty_tail_resume_live_20260418_rerun2 is accepted 10/10, which is the critical proof that replay breadth is now broader than the original flagship chain and that late-session organization analytics no longer depend on ambient chat luck.

• the next transition-authority pass now closes a subtler root-scoped carryover seam inside the shared follow-up path:

  • assistantService.buildAddressFollowupOffer(...) now reads follow-up anchor metadata through the shared continuity helper instead of reconstructing it from yet another local addressDebug parser;
  • assistantTransitionPolicy no longer promotes assistant-side organization authority into previous_anchor_type=organization when a root-scoped inventory pivot intentionally sanitizes the selected-item carryover and keeps only the restored root frame;
  • this matters because root_context_only VAT pivots from inventory drilldown should preserve restored organization/date filters without pretending that restored scope is itself a user-selected follow-up anchor;
  • targeted assistantAddressFollowupContext and assistantTransitionPolicy suites are now green after the fix, explicitly protecting the inventory drilldown -> VAT pivot regression where selected-item carryover must be removed while the inventory root company/date window remains intact;
  • live replay address_truth_harness_phase12_wider_saved_session_pool_live_20260418_rerun10 remains accepted 20/20, which is the critical proof that this anchor-sanitization convergence did not reopen the flagship saved-session continuity path.

• the next replay-breadth pass now proves a different late-session contour around answer inspection and self-correction:

  • a new live pack address_truth_harness_phase15_answer_inspection_followup validates smalltalk -> company fixation -> historical inventory -> selected-item purchase provenance -> selected-item sale trace -> answer inspection -> VAT-on-purchase-date bridge inside one shared session;
  • the first strict replay exposed a real architecture seam rather than a wording issue:
    • after a grounded selected-item sale trace, the user could ask у тебя написано кто контрагент: рабочая станция - это ошибка?;
    • the runtime was still trying to treat that as a fresh address retrieval request, which collapsed into unknown / unsupported instead of inspecting the already grounded previous answer;
  • the fix is now explicit in the orchestration layer:
    • living-mode policy exposes a dedicated answer-inspection signal for self-correction wording;
    • meta follow-up policy can now recognize answer inspection over grounded answer as its own follow-up class instead of leaving it to the generic address lane;
    • route policy now keeps that class out of the address lane and deliberately routes it back into living-chat inspection logic;
    • living-chat runtime now serves a deterministic inspection reply contract for selected-item provenance / sale-trace answers, explicitly distinguishing selected item from counterparty and preserving the next business move;
  • this matters architecturally because another ambient monolith behavior is now an explicit runtime contract:
    • grounded answer inspection is no longer left to accidental prompt luck;
    • self-correction over a previous exact answer can now coexist with selected-object continuity instead of breaking the session into unsupported chat;
    • the neighboring bridge selected-item trace -> VAT on purchase date remains alive after the inspection turn, which proves that answer inspection no longer tears down the active business frame;
  • live replay address_truth_harness_phase15_answer_inspection_followup_live_20260418_rerun5 is accepted 9/9, which is the critical proof that this inspection-follow-up contour now survives as a real saved-session path instead of a one-off manual rescue.

• the next living-chat convergence pass now removes one more local owner from the grounded contextual reply layer:

  • answer-inspection follow-ups are no longer interpreted only inside assistantLivingChatRuntimeAdapter via its own local selected-object reply builder;
  • assistantMemoryRecapPolicy now owns all three grounded contextual deterministic reply classes used by living-chat:
    • inventory-history capability follow-up;
    • memory recap over grounded address context;
    • answer inspection over the previous grounded selected-object answer;
  • resolveAssistantLivingChatMemoryContext(...) now carries explicit contextualAnswerInspectionFollowup plus the grounded address debug that should be inspected, so living-chat reads one shared policy context instead of reconstructing this class from raw mode reason and direct continuity fields inline;
  • this matters because living-chat is now less of a hidden parallel owner of grounded session semantics, and future answer-inspection / recap / capability follow-up fixes can land in one shared policy seam instead of splitting again across adapter-local builders;
  • targeted recap and living-chat runtime tests stay green after this move, and live replay address_truth_harness_phase15_answer_inspection_followup_live_20260418_rerun7 remains accepted 9/9, which is the critical proof that the policy convergence did not reopen the phase15 contour.

• the next living-chat owner-reduction pass now removes one more dead parallel builder layer from the adapter itself:

  • assistantLivingChatRuntimeAdapter no longer keeps local deterministic builders for inventory-history capability, memory recap, or selected-object answer inspection alongside the already active shared policy implementations;
  • those builders were no longer on the active runtime path, but their presence kept a false second owner of the same user-facing behavior inside the adapter and increased the chance that future fixes would land in dead code instead of the shared policy seam;
  • the adapter now imports and uses only the shared builders from assistantMemoryRecapPolicy, which makes the live chat branch structurally closer to a single owner for grounded contextual replies;
  • targeted assistantLivingChatRuntimeAdapter and assistantMemoryRecapPolicy tests stay green after the cleanup, and backend build remains green;
  • live reruns on phase14 and phase15 on 2026-04-19 surfaced partial top-level status only because the packs still pin inventory today expectations to 2026-04-18; the repaired contextual reply contours themselves stayed semantically clean, which confirms this pass as owner reduction rather than a new runtime regression.

• the next continuity-authority pass now removes one more duplicate root-frame owner from assistantService follow-up glue:

  • assistantService.extractAddressCarryoverAnchor(...) no longer reconstructs anchor resolution from raw anchor_value_* / extracted_filters using its own local precedence order;
  • assistantService.findRecentInventoryRootFrame(...) no longer rebuilds inventory root carryover from detected_intent + extracted_filters as a separate local parser;
  • both seams now consume the shared continuity helpers:
    • resolveAddressDebugAnchorContext(...)
    • buildInventoryRootFrameFromAddressDebug(...)
  • this matters because follow-up carryover in the top-level service now reads the same root-frame authority that already owns root_filters / root_anchor / current_frame_kind, instead of keeping a service-local fallback that could silently prefer drilldown extracted_filters over the real address_root_frame_context;
  • targeted assistantAddressFollowupContext and addressInventoryRootFrameRegression suites are green after the move, including a new regression that explicitly proves root_filters come from address_root_frame_context.root_filters rather than from stale drilldown extracted_filters;
  • this pass strengthens continuity convergence in the top-level orchestration glue without introducing a new case-specific branch.

• the next continuity-authority pass now removes one more duplicate carryover owner from assistantTransitionPolicy:

  • transition no longer seeds previous_filters from raw previousAddressDebug.extracted_filters as an isolated local truth source;
  • shared continuity now owns that merge through resolveAddressDebugCarryoverFilters(...), which overlays inventory address_root_frame_context.root_filters onto stale drilldown filters before the follow-up policy starts composing pivots;
  • this matters because the top-level transition glue can now inherit the same root-frame date and warehouse authority that already exists in continuity, instead of silently carrying a stale drilldown as_of_date into root_context_only pivots;
  • targeted assistantContinuityPolicy and assistantTransitionPolicy suites are green after the move, including explicit regression coverage for inventory_purchase_documents_for_item -> inventory_on_hand_as_of_date carryover where root_filters must override a stale drilldown date;
  • this pass reduces one more hidden state-reconstruction fork between the continuity layer and transition glue without introducing case-specific routing;
  • a fresh live rerun of address_truth_harness_phase12_wider_saved_session_pool on 2026-04-19 stayed semantically clean on the repaired carryover path and failed only on the already-known time-unstable today expectations (2026-04-18 vs 2026-04-19) in inventory_root_today, payables_today, and receivables_mirror_today.

• the next continuity-authority pass now centralizes one more shared inventory root-frame seam that used to be split across assistantService and assistantTransitionPolicy:

  • continuity now owns hydrateInventoryRootFrameState(...), which fills missing organization/date scope into inventoryRootFrame and computes currentFrameKind from the same shared state object instead of rebuilding both pieces locally inside transition glue;
  • continuity now also owns buildRootScopedCarryoverFilters(...), so root-scoped filter precedence no longer lives as a separate service-local helper and tests no longer need a legacy re-export from assistantService;
  • this matters because inventoryRootFrame, current_frame_kind, and root-scoped filter precedence now converge through one authority layer before root_context_only pivots are decided, which reduces another hidden chance for state drift when new domains or new follow-up families are added;
  • targeted assistantContinuityPolicy and assistantTransitionPolicy suites are green after the move, with explicit coverage for root-frame hydration from navigation scope and for previous-date precedence over a stale inventory root frame;
  • a fresh live rerun of address_truth_harness_phase12_wider_saved_session_pool on 2026-04-19 remained semantically stable on all repaired continuity paths and again failed only on the already-known date-sensitive today expectations, not on the new shared root-frame state owner.

• the next continuity-authority pass now centralizes temporal backfill precedence for follow-up filters:

  • transition no longer holds a service-local block of shouldBackfillPreviousDateScopeFromNavigation + six field-level ifs for as_of_date / period_from / period_to;
  • shared continuity now owns that merge via applyTemporalCarryoverFilters(...), while shouldUseNavigationTemporalCarryover(...) keeps the intent-family boundary explicit in one place;
  • this matters because navigation date scope and continuity temporal scope are now merged through one owner before transition decides pivots, instead of being backfilled ad hoc inside the hot path;
  • targeted assistantContinuityPolicy and assistantTransitionPolicy suites are green after the move, with direct helper coverage for navigation-first temporal precedence and for non-applicable intent families staying untouched;
  • a fresh live rerun of address_truth_harness_phase12_wider_saved_session_pool on 2026-04-19 stayed semantically stable and again failed only on the already-known date-sensitive today expectations, not on the new shared temporal carryover authority.

• the next continuity-authority pass now centralizes organization carryover precedence for follow-up filters:

  • transition no longer owns a local cascade of historical organization -> shared authority -> continuity snapshot -> navigation organization -> clarification selection;
  • shared continuity now owns that merge through applyOrganizationCarryoverFilters(...), so organization hydration in follow-up filters has a single explicit precedence contract;
  • this matters because previous_filters.organization is now aligned with the same continuity authority story that already drives route, living-chat, and data-scope, instead of keeping one more hot-path-only merge order inside transition glue;
  • targeted assistantContinuityPolicy and assistantTransitionPolicy suites are green after the move, with direct helper coverage for organization precedence and for preserving an already grounded organization value;
  • a fresh live rerun of address_truth_harness_phase12_wider_saved_session_pool on 2026-04-19 remained semantically stable and again failed only on the already-known date-sensitive today expectations, not on the new shared organization carryover authority.

Next Execution Slice (2026-04-18)

The project is now moving from:

• breakpoint recovery

to:

• danger-zone exit under explicit gates

This next slice should be executed in the following order:

1. Finish continuity authority convergence in the hot runtime path.
2. Widen saved-session replay coverage beyond the already repaired flagship chains.
3. Reduce duplicated state reconstruction that still lives outside the shared continuity authority.
4. Prove that non-flagship hybrid/deep-turn contours survive the same runtime contracts across more than one repaired scenario pack.
5. Only after that, continue secondary answer-shaping cleanup where it materially affects acceptance.
6. Only after that, begin controlled domain-by-domain expansion toward the multi-domain stage.

Current explicit goals for this slice:

• fewer owners independently reconstruct active context;
• more replay breadth before any large expansion claim;
• fewer hidden runtime contract mismatches between entry prompts, schema validation, and deep-turn orchestration;
• fewer hybrid/deep entry seams that still depend on fragment luck instead of explicit runtime contracts;
• cleaner user-facing business answers on already-correct truth paths;
• lower risk that new domains multiply orchestration chaos faster than capability growth.

Ready Signal

The project can leave the current breakpoint when:

• mixed live sessions no longer depend on distributed guesswork about active context;
• clarification does not outrank valid restored business continuity;
• recap is grounded and business-useful;
• technical scaffolding is removed from user-facing meta answers;
• the primary mixed replay is green for the real user path, not only for narrow packs.