- Turn start keeps develop's supersede/concurrency, run tracker,
register and forceFinalize, and attaches the TranscriptRedactor.
- The shared finalization flushes redaction after the final events and
before the terminal turn update on every path (supersede, timeout,
restart recovery, forceFinalize); the redactor always closes.
- Token-bearing columns widened to TEXT (H2-safe upgrade statements),
entity length limits removed so tokens never truncate.
- Event reader race that could omit terminal events fixed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The foreign-key block is PL/pgSQL; the H2 loader in
AssistantDefaultsChangelogTest now splits it as one statement and skips it.
SchemaForeignKeyPostgresTest runs it on real Postgres.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Conflicts in CHANGELOG.md and qbit-assistant/pom.xml: kept develop's
durable-turns and starter-defaults entries and its H2/Liquibase test
dependencies alongside the Testcontainers/Postgres ones.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Conflicts in UPGRADE.md and docs/schema-postgres.sql: kept develop's
durable-turns and starter-defaults notes and DDL; the index audit block
now follows the durable-turns block at the end of the schema.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A seeded row cannot be told apart from a host's own row with the same name
('default' profile, policy, starter skill names), and Liquibase also rolls
back changesets it marked ran, so the name-based rollback lines could delete
a host's data. Every seed changeset now rolls back to nothing; the test runs
a real Liquibase rollback over host-owned rows and checks they survive.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Review of #48 found:
- the formatted-SQL table-exists precondition does not parse in Liquibase 5
(and is unsupported in 4.19), failing the host's whole migration: use
sql-checks only;
- the budget and prices were seeded beside a host's own budgets and over its
configured pricing: seed the policy, budget and prices only into empty
tables, the profile only when no default profile exists;
- the Setup card ignored registry and configured prices: price the default
model exactly as calls are costed;
- "Add rows for every table" could read a stale data-access cache.
The changelog test now runs through Liquibase 4.33 (kof22's version):
fresh schema, re-run, an already configured host, and a missing table.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A new question supersedes the running one in the conversation (default) or
runs concurrently; the soft per-user cap supersedes the oldest. Stop works
on any replica and on dead runs, and is idempotent. Orphans are reaped once
their lease lapses (45 s) instead of after maxTurnSeconds + 60 s. 503
SHUTTING_DOWN while a replica stops. turn.phase events and startedAt on
message.start. UI API extension assistant 1.2.0.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A per-turn watchdog (the user's own session) renews the lease, picks up a
Stop requested on another replica, enforces maxTurnSeconds and force-
finalises a wedged engine. One idempotent finaliser: accounting or event
failures can no longer leave a turn RUNNING or registered. On shutdown the
runtime stops taking turns, cancels, waits a grace and finalises the rest
as interrupted. The in-memory concurrent counter is gone. Runs are
published through the AssistantRunTracker stand-in SPI (qqq#1210).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
heartbeat_at, lease_owner, cancel_requested_at, phase and concurrency_mode,
a (user_id, status) index, and a partial unique index so a turn gets one
message.end even when two replicas reap it. Store helpers for a conditional
update of a still-running turn.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Owner requirement 2026-10-14: every assistant table should be populated and
visibly used, and an empty table should say how to set it up.
- db/liquibase/qbit-assistant/assistant-defaults.changelog.sql (host-included,
insert-only, self-guarding): default policy, default profile, four starter
skills linked to it, a $50/month instance budget (warn 80, downgrade 100,
stop 120), current Anthropic prices, a disabled example scheduled task.
- Assistant home Setup card: each config table, its state and the next step.
- assistant_data_access action "Add rows for every table" writes the default
access out as editable rows.
- help/qbit-assistant.yaml (help-bundle format) on the app and every table,
saying what it is for and how rows get created (STAND-IN(core=#1198)).
- Tests: guardrail (every table used at runtime, seeded or explained, with
empty-state help); changelog seeds a fresh H2 schema insert-only; seeded
rows drive persona, skills in the model request, budget warning, policy.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Reference DDL gains a guarded, re-runnable foreign key block after the
index audit: the conversation tree CASCADEs (matching
AssistantStore.deleteConversation and retention), the message/turn cycle,
usage and the data audit are SET NULL so cost rows and the audit trail
survive, provider references from usage, budget accounting and the data
audit are NO ACTION, profile-skill links CASCADE, and
assistant_model.provider_id becomes CASCADE (fk_assistant_model_provider,
as on kof22). Orphans are recorded in qqq_fk_quarantine first. Every FK
column now leads an index.
SchemaForeignKeyPostgresTest applies the DDL on Postgres (Testcontainers)
and fails any declared relation without a foreign key, a wrong delete
rule, an unindexed FK column, or a broken conversation delete order.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The reference Postgres schema gains six IF NOT EXISTS indexes:
- assistant_usage (turn_id): ScheduledTurn.cost
- assistant_turn_event (create_date), assistant_tool_call (create_date) and
assistant_conversation (modify_date): the retention filters
- assistant_scheduled_run (conversation_id) and (turn_id): the ON DELETE
SET NULL foreign keys
UPGRADE.md notes them.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Stored copies of tool results replace PII/PHI field values with
ClassifiedValueTokens.forRecordField tokens (core, QRun-IO/qqq#1182);
model-written text, persisted text deltas, presentation parts, tool
inputs, run summaries and audit reasons have every classified value
the turn saw replaced by its token. What the model receives (as the
PHI routing allows) and the live SSE to the entitled user are
unchanged.
- Pending writes that would carry a classified value are refused rather
than stored (follow-up: an encrypted short-lived payload).
- A core log redaction hook masks seen classified values in assistant
logs.
- Canary test: CANARY-PHI/PII markers never appear (nor any >4-char
substring) in any assistant_* table, logs or audits, while the model
did receive them in memory.
- Requires QQQ 4.1.0-RC.7.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The qbit-assistant-provider-api artifact now holds exactly one package,
com.kingsrook.qbits.assistant.spi (23 types including ModelPricing), so
consumers such as qbit-marketing depend only on that artifact and the
qbit-bom G1 guardrails can allow it by package. BREAKING: imports change
from ...assistant.provider.X / ...assistant.engine.ModelPricing to
...assistant.spi.X (UPGRADE.md has the table). ArchUnit forbids split
packages between the API and the implementation.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Moves the optional email integration from qbit-email internals to
com.kingsrook.qbits.email.api (qbit-email 0.1.0-RC.2) with identical
behaviour, and moves to qbit-build-parent 2.1.1, whose default G1
allow-lists include qbit-email's API (QRun-IO/qbit-bom#12), so the
per-repo allow-lists are gone.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Moves qqq-orb@0.6.8 to qqq-orb@0.6.9, which passes --update-snapshots so cached SNAPSHOT dependencies are re-checked on every build (QRun-IO/qqq-orb#22).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Default off (owner: hard wall; kof22 does not opt in). When on, calls
core's ClientChannelsMetaData.withAssistantAuditReads() (QRun-IO/qqq#1193:
audits readable, never writable, worm_audit_* still denied) through a
STAND-IN(core=QRun-IO/qqq#1193, removeBy=RC.8) lookup, falling back on
RC.6 to removing exactly the audit* pattern. Other patterns are never
touched.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Keeps develop's changes and re-applies the optional qbit-email 0.1.0-RC.1
delivery on top (adapted to develop's refactors; guardrail allow-lists
for the optional integration where the build parent enforces them).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
X1:
- Every assistant_* table, process, report and widget (and every object
the qbit produces) is denied to client channel assistant: the
assistant_* pattern in ClientChannelsMetaData plus
QPermissionRules.withDeniedClients on each object.
- Startup fails if any of them is not denied.
- Every UI API call carries assistant/<turnId> (scheduled:
assistant/scheduled/<taskId>); confirm and explain-access carry the
owning turn.
- Tests: assistant.admin through the assistant is denied (403, absent
from list_tables and /metaData) on every configuration object, while the
same admin through next-ui is allowed; the startup check fails on one
undenied object; core's security-surface patterns are present.
Y1:
- assistant_tool_call.audit_href: an /app link to the written record's
audit rows filtered by clientChannel, built with the shared filter-link
builder.
- Tests: an assistant write audits the human actor with
clientChannel assistant/<turnId>; the same write from the UI audits
next-ui. Hosts enable the audit clientChannel field (documented).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Moves from 0.1.0-SNAPSHOT to the first release candidate. Test schemas
gain the RC's email_provider_sender_address table and the message
from_address_id column where they copy the qbit's DDL.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Scheduled runs go through the invoker as assistant/scheduled/<taskId>;
the designated-user manageSession login stays a plain UI-client login.
get_record_history is offered only when the host opts the audit tables
in (core's default wall denies audit* to the assistant). Reserved-header
filtering in the stand-in header invoker is tested.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- CoreUiApiGateway on UiApiInvoker for every tool (P tools, bulk-load,
presentation), client channel assistant/<turnId>; form params mapped to
withFormParam; LoopbackUiApiGateway deleted.
- Extra headers (X-QQQ-AI-Route) through HeaderedUiApiInvoker, a stand-in
until core UiApiRequest has withHeader.
- Core DataClassification replaces the qbit stand-in; the default
classification source merges core metadata with the config map.
- Bulk-load confirm dispatch kept on the spec'd confirm path.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Phase (d) of QRun-IO/qbit-email docs/DESIGN.md section 11. qbit-email is an
optional dependency (0.1.0-SNAPSHOT until its first RC) used through one
adapter class behind a classpath and instance guard; without it the
existing behaviour is unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The owner asked (2026-10-13) that the assistant use and create bulk load
profiles. This adds the spec #1171 section 3.4 tools, all through the
/qqq/v1 UI API with the user's own session:
- bulk_profiles_list / bulk_profile_describe: the user's and shared
profiles, read through the savedBulkLoadProfile table (record locks).
- bulk_file_attach: an upload the panel already made (pageContext
attachments, refs only), with its masked column profiles.
- bulk_mapping_suggest: core's suggester. The model then refines the
profile from the headers and the permitted samples.
- bulk_load_dry_run: the table's real bulk-insert validation (SAMPLE or
FULL), with nothing stored. Error groups come with row numbers, so the
model can fix the profile and run again.
- bulk_profile_save (write): storeSavedBulkLoadProfile, plus
insertSharedRecord for each share. It runs only after the user confirms
the pending action, and only as far as the user may share.
- bulk_load_run (write): goes through a BulkLoadRunner seam after
confirmation. The default runner is the durable jobs launch, where core
creates assistant jobs AWAITING_CONFIRMATION. Until that ships, it
answers that the launch is not available yet (STAND-IN).
PHI: the X-QQQ-AI-Route header (hipaa|standard) comes from the turn's
provider HIPAA eligibility, never the model. The server masks samples and
values. Tool-call logs drop sampleValues, example bullets and value
mappings. Data access READ_WRITE is required to propose and to confirm
writes.
Supporting changes, all additive:
- UiApiCall: form params and headers.
- LoopbackUiApiGateway: multipart bodies.
- UiApiToolbox.Outcome.of.
- One dispatch branch each in ContextBuilder, TurnEngine and
AssistantApi.confirmAction.
- ToolLabels entries.
- The config seam.
The core operations are QRun-IO/qqq#1179 (STAND-IN until an RC). Until
then the tools say the server does not offer them.
Tests: BulkLoadToolsTest (6), including the golden eval. It uploads a CSV
with messy headers. The assistant proposes a profile, the dry run finds
the missing required field, the assistant defaults it, the dry run
passes, it saves and shares the profile after confirmation, and in a
second turn runs the load after confirmation. Full ./mvnw verify is green.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
james.maesandClaude Opus 5.5