ltcai 8.9.0 → 9.0.0

package/README.md

@@ -1,6 +1,6 @@
 # Lattice AI
 
-**Lattice AI 8.9.0 is the local-first Digital Brain platform. This release hardens authenticated history and graph scoping, blocks unapproved direct Tool API execution paths, makes local permission approvals and installer/process execution safer to audit, and keeps package, desktop, extension, static app, and documentation metadata synchronized to the 8.9.0 line.**
+**Lattice AI 9.0.0 is the local-first Digital Brain platform. This release closes the July 8 code-review follow-up, hardens chat/runtime reliability, pays down cleanup debt across runtime utilities and setup detection, and keeps package, desktop, extension, static app, and documentation metadata synchronized to the 9.0.0 line.**
 
 **Lattice AI는 모델이 바뀌어도 내 지식과 맥락을 보존하는 로컬 우선 AI 브레인입니다.**
 
@@ -67,9 +67,9 @@ You need Lattice AI when:
 The screenshots below are the latest checked-in visual evidence captures. They
 keep the first-run Brain flow, memory graph, source capture, model library,
 system view, admin console, and review center visible as release gates while
-8.9.0 focuses on scoped memory isolation, direct Tool API policy enforcement,
-AgentRuntime human-approval semantics, confirmation-token guarded installer
-execution, and frontend/runtime maintainability.
+9.0.0 focuses on code-review closure, chat/runtime reliability,
+AgentRuntime and ToolRegistry maintainability, release artifact correctness,
+and frontend/runtime cleanup.
 
 ### 1. Wake Brain
 
@@ -210,40 +210,38 @@ See [docs/DEVELOPMENT.md](docs/DEVELOPMENT.md) for developer workflow details.
 
 ## Current Release
 
-The current release is **8.9.0 — Scoped Memory & Tool Policy Hardening**:
-
-- Conversation history reads and deletes now scope by authenticated user and
-  readable workspace, while preserving legacy-global rows only in no-auth or
-  explicitly compatible flows.
-- Knowledge Graph search, graph traversal, relationships, node reads, and chat
-  context now apply workspace scope at the graph/service boundary.
-- Direct HTTP/MCP Tool API calls enforce ToolRegistry policy before hooks or
-  handlers run; destructive paths and non-auto-approved user writes are blocked.
-- AgentRuntime no longer treats “requires approval” plans as implicitly
-  approved, and rollback now handles tool results that omit `success`.
-- Local permission approvals hash tokens at rest, use atomic queue writes, and
-  reject write approvals for blocked system prefixes.
-- Frontend API base handling, Tauri credentials, workspace clearing, CSS tokens,
-  and i18n literal checks are split into safer maintainability seams.
-
-Expected artifacts for 8.9.0 release must use exact filenames:
-
-- `dist/ltcai-8.9.0-py3-none-any.whl`
-- `dist/ltcai-8.9.0.tar.gz`
-- `ltcai-8.9.0.tgz`
-- `dist/ltcai-8.9.0.vsix`
-- `src-tauri/target/release/bundle/dmg/Lattice AI_8.9.0_aarch64.dmg`
+The current release is **9.0.0 — Code Review Closure & Runtime Cleanup**:
+
+- July 8 code-review follow-ups are fixed across chat, agent runs, permissions,
+  Telegram integration, runtime audit timestamps, embeddings, and Brain UI copy.
+- Chat/document streaming now preserves terminal SSE events and history/trace
+  persistence when generation fails mid-stream.
+- Runtime audit appends use JSONL while legacy JSON reads remain compatible.
+- Shared JSON/ISO/hash, setup detection, and frontend helper utilities replace
+  duplicated implementations across runtime, Brain Core, setup, and React code.
+- The legacy `server_app` namespace is allowlist-based instead of exposing all
+  app-factory locals.
+- The main `/chat` handler now routes through focused intent handlers and shared
+  response epilogues for lower future regression risk.
+
+Expected artifacts for 9.0.0 release must use exact filenames:
+
+- `dist/ltcai-9.0.0-py3-none-any.whl`
+- `dist/ltcai-9.0.0.tar.gz`
+- `ltcai-9.0.0.tgz`
+- `dist/ltcai-9.0.0.vsix`
+- `src-tauri/target/release/bundle/dmg/Lattice AI_9.0.0_aarch64.dmg`
 
 Do not use wildcard artifact uploads. Package registry publishing remains owner-run.
 
 See [docs/ROADMAP_RECOMMENDATIONS.md](docs/ROADMAP_RECOMMENDATIONS.md) for the
-strategic roadmap slices applied through 8.9.0 and the follow-up tracks.
+strategic roadmap slices applied through 9.0.0 and the follow-up tracks.
 
 ## Known Limitations
 
 - External package registries are owner-published and can lag behind GitHub.
 - PostgreSQL/pgvector is optional scale/migration tooling. SQLite remains the
-  live local Brain store in 8.9.0.
+  live local Brain store in 9.0.0.
 - Docker, model downloads, cloud model calls, Telegram, Brain Network, and update
   checks require explicit user action.
 - Conversation does not fabricate answers when no model is loaded.
@@ -255,6 +253,7 @@ strategic roadmap slices applied through 8.9.0 and the follow-up tracks.
 
 | Version | Theme |
 | --- | --- |
+| 9.0.0 | Code Review Closure & Runtime Cleanup: July 8 code-review follow-ups fixed, chat/runtime reliability improved, duplicated utility surfaces consolidated, runtime audit append paths moved to JSONL, and release metadata/artifacts synchronized |
 | 8.9.0 | Scoped Memory & Tool Policy Hardening: authenticated history/KG reads are workspace-scoped, direct Tool API paths enforce registry policy, local approvals hash tokens at rest, AgentRuntime approval semantics are explicit, and frontend/runtime seams are split |
 | 8.8.0 | Brain Core Extraction & Recall Proof Hardening: internal-only Brain shim layers are removed, AgentRuntime run contracts/retry budgets are tighter, Brain Chat gains conversation controls, and citation recall exposes matched evidence |
 | 8.7.0 | Runtime State Hygiene & Release Evidence Refresh: model-runtime internals prefer typed state over legacy globals, compatibility sync is deprecated, 8.7.0 visual evidence is refreshed, and all release metadata/docs are synchronized |
@@ -265,16 +264,6 @@ strategic roadmap slices applied through 8.9.0 and the follow-up tracks.
 | 8.2.0 | Brain Brief: evidence-backed home briefing, honest empty-state guidance, recall/graph/model-proof next actions, and continued model/workspace runtime extraction |
 | 8.1.0 | Intuitive Brain Home: living Brain, recent memory, connected topic, next action, and composer are visible in one product-first screen with refreshed 8.1.0 evidence and artifacts |
 | 8.0.0 | Runtime Architecture Contract: AgentRuntime, ToolRegistry, central Config, server decomposition, and KG hardening are captured as machine-checkable release boundaries with exact 8.0.0 artifacts |
-| 7.9.0 | Agent Runtime Boundary Hardening: explicit `SingleAgentRuntime`, compatibility alias preservation, injected rollback port, and release/readiness docs aligned to the product AgentRuntime facade |
-| 7.8.0 | Brain Chat Home UX Simplification: chat-first first viewport, visible workspace navigation, collapsed source/status utilities, hidden default depth controls, and removal of obsolete Brain UX components |
-| 7.7.0 | Complete Product Polish: command-center Brain Home, repaired Review Center evidence, exact 7.7 docs/artifacts, product readiness gate, and stronger CI/release checks |
-| 7.6.0 | Brain-Centered UX & Architecture Closure: Wake Brain first-run surface, concentric memory rings with direct depth controls, and machine-checkable closure of the two architecture/UX review files |
-| 7.5.0 | Runtime Debt Burn-down & Release Risk Cleanup: API consumers get normalized contract views, retrieval quality uses a 250+ record corpus fixture, stale artifact risk is removed, npm audit is clean, and Tauri is updated past the old block warning |
-| 7.4.0 | Runtime Contract Convergence & Corpus Retrieval: agent/workflow/audit/realtime records share the agent-run-contract/v1 family, and retrieval quality gates run against a real corpus-scale fixture |
-| 7.3.0 | Runtime Contract & Retrieval Quality: shared agent-run contract across runtimes plus deterministic hybrid recall/ranking regression gates |
-| 7.2.0 | Runtime Trust Baseline: agent run preview/readiness, simulation-mode guardrails, live ToolRegistry manifest/diagnostics, and tests for dispatch/governance/catalog drift |
-| 7.1.0 | Brain Usability Completion: clearer first-run onboarding, ingestion progress/emergence, richer graph controls, inline answer proof, workspace/profile/admin discovery, empty/error/consent feedback, and VS Code sync status |
-| 7.0.0 | Brain Productization Loop: first-screen ingestion for files/folders/notes/web, answer-level memory proof and source citations, model-continuity demo flow, five-minute first-run loop, and recall/KG quality eval in CI |
 
 ## Documentation
 

package/auto_setup.py

@@ -54,6 +54,12 @@ from latticeai.services.process_audit import (
     command_plan_for_commands,
     require_command_confirmation,
 )
+from latticeai.services.setup_detection import (
+    detect_cuda,
+    detect_tools,
+    detect_wsl_from_text,
+    parse_windows_video_controllers as _parse_windows_video_controllers,
+)
 
 __all__ = [
     "SystemProfile", "Recommendation", "InstallPlan",
@@ -153,44 +159,6 @@ def _which(binary: str) -> Optional[str]:
     return None
 
 
-def _parse_windows_video_controllers(raw: str) -> List[Dict[str, Any]]:
-    controllers: List[Dict[str, Any]] = []
-    if not raw:
-        return controllers
-    try:
-        data = json.loads(raw)
-        if isinstance(data, dict):
-            data = [data]
-        if isinstance(data, list):
-            for item in data:
-                name = str(item.get("Name") or "").strip()
-                if not name:
-                    continue
-                try:
-                    ram_mb = int(item.get("AdapterRAM") or 0) // (1024 * 1024)
-                except Exception:
-                    ram_mb = 0
-                controllers.append({"name": name, "vram_mb": ram_mb})
-        if controllers:
-            return controllers
-    except Exception:
-        pass
-    current: Dict[str, Any] = {}
-    for line in raw.splitlines():
-        if line.startswith("Name="):
-            if current:
-                controllers.append(current)
-            current = {"name": line.split("=", 1)[-1].strip(), "vram_mb": 0}
-        elif line.startswith("AdapterRAM=") and current:
-            try:
-                current["vram_mb"] = int(line.split("=", 1)[-1].strip()) // (1024 * 1024)
-            except ValueError:
-                current["vram_mb"] = 0
-    if current:
-        controllers.append(current)
-    return controllers
-
-
 def _detect_gpu(prof_os: str, arch: str) -> GPUInfo:
     """OS별 휴리스틱으로 GPU 감지. 외부 라이브러리 없이 가능한 만큼만."""
     gpu = GPUInfo()
@@ -276,36 +244,17 @@ def _detect_package_manager(prof_os: str) -> Optional[str]:
 
 
 def _detect_tools() -> Dict[str, str]:
-    tools: Dict[str, str] = {}
-    for binary in ("ollama", "lms", "nvidia-smi", "nvcc", "winget", "brew", "apt", "git", "node", "python", "python3"):
-        found = _which(binary)
-        if found:
-            tools[binary] = found
-    return tools
+    detected = detect_tools(_which, ("ollama", "lms", "nvidia-smi", "nvcc", "winget", "brew", "apt", "git", "node", "python", "python3"))
+    return {binary: path for binary, path in detected.items() if path}
 
 
 def _detect_wsl(prof_os: str) -> Tuple[bool, str]:
-    if prof_os != "linux":
-        return False, ""
-    raw = _read_text("/proc/version")
-    is_wsl = "microsoft" in raw.lower() or "wsl" in raw.lower()
-    version = "2" if "microsoft-standard" in raw.lower() or "wsl2" in raw.lower() else ("1" if is_wsl else "")
-    return is_wsl, version
+    return detect_wsl_from_text(prof_os, _read_text("/proc/version"))
 
 
 def _detect_cuda() -> Tuple[bool, str]:
-    nvidia_smi = _which("nvidia-smi")
-    nvcc = _which("nvcc")
-    version = ""
-    if nvidia_smi:
-        raw = _run([nvidia_smi, "--query-gpu=driver_version", "--format=csv,noheader"], timeout=4.0)
-        version = raw.splitlines()[0].strip() if raw.splitlines() else ""
-    if nvcc:
-        raw = _run([nvcc, "--version"], timeout=4.0)
-        m = re.search(r"release\s+([\d.]+)", raw)
-        if m:
-            version = m.group(1)
-    return bool(nvidia_smi or nvcc), version
+    available, version, _nvidia_smi, _nvcc = detect_cuda(_which, lambda args: _run(args, timeout=4.0))
+    return available, version
 
 
 def _detect_cpu_details(prof_os: str) -> Tuple[str, int, int, List[str]]:

package/docs/CHANGELOG.md

@@ -6,6 +6,81 @@ existed at that release.
 
 ## [Unreleased]
 
+No unreleased changes yet.
+
+## [9.0.0] - 2026-07-08
+
+### Added
+- Added Brain Brief suggested questions that turn current memory, recall proof,
+  graph concepts, and conversation history into clickable first-screen prompts.
+- Suggested Brain questions now send immediately from the first screen instead
+  of only filling the composer.
+- Added one-click follow-up prompts under the latest Brain answer for turning a
+  reply into a checklist, evidence review, or prioritized next steps.
+- Added a Brain chat to Review Center handoff so users can save an answer as a
+  reviewable task draft and manage it alongside automation suggestions.
+- Added direct Brain-to-Agent delegation and successful agent-run synthesis into
+  durable Brain memory/graph context.
+- Surfaced recent agent-synthesis memories in Brain overview and memory rings
+  so delegated work is visibly reflected on the home screen.
+- Improved agent-run Brain synthesis quality by splitting successful results
+  into key facts, decisions, and follow-up memories with structured metadata.
+- Agent follow-ups now enter Review Center as task drafts so delegated work
+  produces actionable approval candidates instead of passive memory only.
+- Approving an Agent follow-up review item now promotes it into a manual
+  workflow draft with trigger, agent, and output nodes.
+- Added large-feature foundations for KG/Retrieval scale diagnostics,
+  background ingestion scheduling, offline multimodal image captions,
+  proactive contradiction detection, and ingestion bridge marketplace templates.
+- Added proactive Brain action cards that turn Brain Brief evidence into
+  one-click ask, Agent delegation, Review Center draft, or graph navigation
+  actions on the Brain home screen.
+- Added a visible proactive Brain action trail so one-click suggestions show
+  their running/completed/failed state after the user acts on them.
+
+### Changed
+- Continued app-factory decomposition by extracting user profile/API-key helper
+  wiring into `latticeai.runtime.user_key_runtime`, keeping the legacy
+  `server_app` callable surface while making keyring/plaintext fallback policy
+  independently testable.
+- Split additional runtime and static-data seams out of app, chat, MCP, model,
+  and Knowledge Graph modules while preserving re-export compatibility for
+  existing imports.
+- Routed Computer Use direct `/cu/*` actions through the shared ToolRegistry
+  policy gate and audit lifecycle, preserving route paths while blocking
+  non-admin direct desktop control by default.
+- Moved blocked-system-prefix protection into `tools.local_write` itself so
+  local filesystem writes fail closed even when called outside the HTTP approval
+  route.
+- Narrowed the lazy `server_app` compatibility namespace by filtering
+  app-factory scratch imports and runtime wiring dictionaries while preserving
+  explicit legacy helpers, and added a typed `RuntimeBundle` migration target
+  behind the legacy `_RUNTIME_BUNDLE` dict.
+
+### Fixed
+- Added regression coverage for provider API-key lookup/storage behavior,
+  including keyring precedence, plaintext fallback gating, legacy plaintext
+  cleanup after keyring writes, and identity creation on plaintext fallback.
+- Added regression coverage for Computer Use policy enforcement, audit-safe
+  typed-text metadata, and direct local-write system-prefix blocking.
+- Fixed functional findings from the July 8 code review: file generation now
+  fails cleanly when no model is loaded, chat/document streams preserve a
+  terminal SSE event on generation errors, agent runs persist failed status on
+  executor exceptions, Brain delegation treats HTTP failures as failures, and
+  local permission expiry cleanup no longer corrupts the active token lookup.
+- Tightened non-security chat intent detection, Telegram bot server URL
+  configuration, LATTICE_TZ-aware runtime audit timestamps, local embedding
+  dimension consistency, and stale Brain UI version copy.
+- Paid down the remaining July 8 cleanup debt by moving duplicated JSON/ISO/hash
+  and setup detection helpers into shared modules, switching runtime audit
+  appends to JSONL while preserving legacy JSON reads, making the legacy runtime
+  namespace allowlist-based, clarifying static-vs-SPA design token ownership,
+  and consolidating duplicated frontend helper functions.
+- Reduced the remaining chat-router risk by extracting repeated chat history,
+  bridge notification, no-model, single-answer, direct-file, and agent-file
+  response paths out of the main `/chat` handler, with regression coverage for
+  the shared fast-path epilogue.
+
 ## [8.9.0] - 2026-07-06
 
 ### Added
@@ -233,7 +308,7 @@ existed at that release.
 - Updated architecture and product readiness targets to 8.0.0.
 - Synchronized package/runtime/static/Tauri metadata to 8.0.0.
 - Updated current-release docs and exact artifact names to 8.0.0 while
-  preserving historical 7.x entries.
+  setting 8.0.0 as the oldest retained release-history entry.
 
 ### Fixed
 - Made logical Knowledge Graph `replace` imports transactional so malformed
@@ -242,239 +317,3 @@ existed at that release.
   `relationship_search`, and `traverse`.
 - Preserved colliding legacy edge labels during logical import/backfill while
   keeping native write-door synonym dedupe canonical.
-
-## [7.9.0] - 2026-06-23
-
-### Changed
-- Added `SingleAgentRuntime` as the explicit name for the legacy single-agent
-  state machine while preserving `AgentRuntime` as a compatibility alias.
-- Updated tool dispatch to build `SingleAgentRuntime` directly.
-- Moved single-agent git rollback behind an injected `rollback_file` port owned
-  by `ToolDispatchService`, keeping shell execution out of the core state
-  machine.
-- Added a shared `runtime-boundary/v1` descriptor so product and single-agent
-  runtime boundaries are machine-readable in config/tests.
-- Added `RuntimeBoundaryProtocol` as the minimal shared inspection surface for
-  runtime-boundary-aware dependency injection.
-- Updated architecture and product readiness targets to 7.9.0.
-
-## [7.8.0] - 2026-06-22
-
-### Changed
-- Rebuilt Brain Chat Home around immediate conversation: chat purpose, starter
-  prompts, and the composer now occupy the first viewport.
-- Collapsed source ingestion, readiness, proof, timeline, overview, model
-  continuity, and care controls behind one utility drawer.
-- Kept workspace navigation visible on the default Brain surface.
-- Hid default depth controls until the user intentionally travels deeper.
-- Integrated the six post-7.7 UX drafts into the canonical Brain experience:
-  first-run value cards, stronger recommendation affordances, product-toned
-  Brain Home copy, and routed legacy Brain conversation entry points to the
-  canonical Brain Home surface.
-- Removed obsolete Brain conversation and first-run guide components.
-- Moved draft onboarding polish out of inline styles and into the shared design
-  stylesheet with responsive behavior and bilingual copy.
-- Updated architecture and product readiness targets to 7.8.0.
-- Refreshed 7.8.0 release screenshots, walkthrough video/GIF, and capture
-  notes under `output/release/v7.8.0/`.
-
-## [7.7.0] - 2026-06-22
-
-> 7.7.0 marks the complete, finished product stage for Lattice AI.
-> After 7.6.0 architecture closure, this release polishes every surface so that anyone looking at the code, UI, docs, or running app immediately recognizes: "this is now a product".
-
-Lattice AI v7.7 delivers the Living Brain as the undeniable center, production-grade runtime contracts, stable ToolRegistry, full ingestion-to-graph flows, bilingual professional UX, and zero-beta signals. Classifiers moved to Production/Stable. All prior gates remain enforced under finished product contract.
-
-Package metadata, Tauri, frontend, Python all aligned to 7.7.0. UI/UX microcopy and signals updated to convey finished professional tool.
-
-### Productization Highlights
-- Extreme self + claude-code (pts_claudecode) used for polish, evaluation, iteration.
-- "This is a product" bar: clear durable knowledge ownership, no loose ends.
-- Validation: typecheck, unit, cargo, build scripts exercised.
-
-### Changed
-- Package/runtime/static metadata synchronized to 7.7.0.
-- Development status to Production/Stable.
-- All current-release references point to 7.7.0.
-
-## [7.6.0] - 2026-06-22
-
-> Brain-Centered UX & Architecture Closure. Incorporates the two local review
-> files into the release line with a Wake Brain first-run surface, memory rings
-> plus direct depth controls, and machine-checkable architecture readiness gates.
-
-### Added
-- Wake Brain first-run entry before owner/profile setup, reducing onboarding to
-  the product promise first and the setup mechanics second.
-- Concentric memory rings around Brain Home plus direct controls for Now, Memory,
-  Topics, Relationships, and Full Graph navigation.
-- `latticeai.services.architecture_readiness.architecture_readiness()` and
-  `tests/unit/test_v76_review_completion.py` to keep AgentRuntime, ToolRegistry,
-  Config, server decomposition, KG hardening, and Brain UX review closure under
-  test.
-
-### Changed
-- Package/runtime/static metadata is synchronized to 7.6.0; package publish and
-  deployment remain owner-run only.
-- README and release docs now describe 7.6.0 as the current release and point to
-  refreshed 7.6.0 screenshots, walkthrough GIF, and release evidence index under
-  `output/release/v7.6.0/`.
-
-## [7.5.0] - 2026-06-20
-
-> Runtime Debt Burn-down & Release Risk Cleanup. Turns the 7.4.0 contract
-> envelope into a consumed API surface, expands retrieval quality to a 250+
-> record local corpus fixture, and removes release/security warnings.
-
-### Added
-- `extract_contract`, `require_contract`, `contract_view`, and `contract_views`
-  helpers for consumers that need a surface-agnostic `agent-run-contract/v1`
-  projection.
-- AgentRuntime status/list/detail/events and realtime feed responses now expose
-  compact `contracts` views alongside legacy payloads.
-- Deterministic 250+ record retrieval benchmark corpus while keeping 12 judged
-  queries and real `KnowledgeGraphStore` + `SearchService` execution.
-- Refreshed README release evidence screenshots and walkthrough GIF under
-  `output/release/v7.5.0/`.
-
-### Changed
-- Tauri Rust/CLI dependencies are updated within the Tauri 2 line, removing the
-  old transitive `block v0.1.6` future-incompatibility warning.
-- npm dependency overrides move `js-yaml` to a non-vulnerable version; `npm
-  audit` reports 0 vulnerabilities.
-- CI lint compatibility is restored for the Brain quality gate script.
-- Local MLX model preparation now recognizes valid existing Hugging Face cache
-  snapshots, avoiding an unnecessary re-download when the model already exists
-  outside Lattice's managed `~/.ltcai/hf-models` directory.
-- Package/runtime/static metadata is synchronized to 7.5.0; package publish and
-  deployment remain owner-run only.
-
-## [7.4.0] - 2026-06-20
-
-> Runtime Contract Convergence & Corpus Retrieval. Completes the
-> agent-run-contract/v1 family across run storage, workflow execution, audit
-> events, realtime events, and a real corpus-scale retrieval quality gate.
-
-### Added
-- Persisted agent and workflow run rows now carry refreshed contract metadata
-  for queued, running, terminal, cancelled, and interrupted states.
-- Workflow engine results, replay payloads, audit log events, and realtime SSE
-  feed events now expose the same `agent-run-contract/v1` family envelope while
-  preserving existing top-level compatibility fields.
-- Corpus-scale retrieval fixture with 30+ documents, judged queries,
-  must-include expectations, and thresholds for recall, precision, NDCG, and
-  hit rate.
-- `scripts/brain_quality_eval.py` now exercises the real local
-  `KnowledgeGraphStore` + `SearchService` hybrid retrieval path before scoring.
-
-### Changed
-- `RetrievalBenchmarkRunner` reports dynamic metric aliases for the selected
-  `top_k` and a `must_include_hit_rate`.
-- Package/runtime/static metadata is synchronized to 7.4.0; package publish and
-  deployment remain owner-run only.
-
-## [7.3.0] - 2026-06-20
-
-> Runtime Contract & Retrieval Quality. Turns the next AgentRuntime extraction
-> step and the uploaded roadmap's hybrid-search quality goals into a small,
-> tested release: shared run contracts and deterministic recall regression.
-
-### Added
-- `lattice_brain.runtime.contracts.AgentRunContract`, a serializable
-  `agent-run-contract/v1` payload shared by single-agent and multi-agent
-  execution paths.
-- Multi-agent API result/run patches now include the shared contract with
-  runtime, mode, status, roles, retries, timeline, and terminal-state data.
-- Single-agent runtime exposes the same contract helper for UI/API/storage
-  convergence in the next extraction pass.
-- `scripts/brain_quality_eval.py` now runs deterministic hybrid recall/ranking
-  regression checks with recall and precision thresholds.
-
-### Changed
-- Package/runtime/static metadata is synchronized to 7.3.0; package publish and
-  deployment remain owner-run only.
-
-## [7.2.0] - 2026-06-20
-
-> Runtime Trust Baseline. Adds execution preview and registry diagnostics so
-> agent runs and tool permissions become inspectable contracts before action.
-
-### Added
-- `POST /agents/api/run/preview` for AgentRuntime readiness, role selection,
-  input keys, retry clamping, and blocking reasons without starting a run.
-- `GET /tools/registry` for the live ToolRegistry manifest across dispatch
-  handlers, governance policy, catalog descriptions, and permissions.
-- `GET /tools/registry/diagnostics` for a compact drift check suitable for CI,
-  admin views, and runtime health panels.
-- Unit coverage for AgentRuntime preview and ToolRegistry manifest diagnostics.
-
-### Changed
-- Tool governance now covers `read_document`, and the catalog describes
-  `create_web_project`, closing the current dispatch/governance/catalog drift.
-- Package/runtime/static metadata is synchronized to 7.2.0; package publish and
-  deployment remain owner-run only.
-
-## [7.1.0] - 2026-06-20
-
-> Brain Usability Completion. Completes the 7.1.0 first-run through editor-sync
-> usability pass: clear onboarding, visible ingestion, graph controls, inline
-> proof, workspace/admin discovery, feedback states, and VS Code sync status.
-
-### Added
-- Hardware visualization, expected timing, install timeline, and next-action
-  copy in first-run onboarding.
-- Brain Home ingestion stage disclosure and memory emergence timeline for file,
-  folder, note, and URL sources.
-- Knowledge Graph search suggestions, entity type filters, recent/all-time time
-  exploration, focus clearing, and neighbor highlighting.
-- Inline answer citation markers with keyboard-accessible proof cards.
-- Workspace/profile switcher, Admin Console gate, consent revoke feedback, and
-  shared empty/error feedback surfaces in the Brain shell.
-- VS Code extension heartbeat/status endpoint plus extension and main-app sync
-  indicators for connected/indexing/synced/offline states.
-
-### Changed
-- Package/runtime/static metadata is synchronized to 7.1.0; package publish and
-  deployment remain owner-run only.
-
-## [7.0.0] - 2026-06-18
-
-> Brain Productization Loop. Turns the Brain proof work into a first-five-minute
-> product flow: add sources, ask a question, see proof/citations, and verify the
-> same Brain evidence after switching models.
-
-### Added
-- Brain Home ingestion panel for files, local folder paths, notes, and web URLs,
-  all backed by existing workspace-scoped ingestion routes.
-- Answer-level Memory proof and source citation cards rendered under assistant
-  responses after Brain proof refreshes for the user's query.
-- Model-continuity demo strip that lets the user recheck the same Brain
-  evidence and jump to model switching from the Brain flow.
-- Deterministic `scripts/brain_quality_eval.py` recall/KG quality gate, wired
-  into CI after the unit suite.
-- Visual mock coverage for Brain proof, document upload, note ingest, folder
-  indexing, and web URL ingestion endpoints.
-
-### Changed
-- Brain Home is now ingestion-first instead of chat-first: first screen action
-  labels are files/folders/notes/web, with deeper graph/model/settings still
-  reachable from the shell.
-- Package/runtime/static metadata is synchronized to 7.0.0; package publish and
-  deployment remain owner-run only.
-
-- Completed ALL Recommended next refactor items from report #15 in this session:
-  - Server decomp wave (model_runtime globals/wiring): model_loading.py for prepare_and_load and stream, _MODEL_RUNTIME_STATE, model_engines.
-  - Deeper WorkspaceOSStore (timeline + plugins + snapshots + memory).
-  - KG embed: set_embed_dim.
-- All refactoring needed finished this session per AGENTS. 767 tests, builds, greps clean.
-
-### Added (this session features)
-- `vision_analyze` tool: new multimodal vision analysis tool using screenshot b64 + prompt. Leverages existing VLM support (image_data in generate). Added to computer-use agent prompt and general tools. Fits seamlessly with computer_use, agent runtime, tool registry, and VLM models without affecting text-only paths.
-- More recent multimodal models in user recommendations (Llama 3.2 11B Vision, Phi-3.5 Vision, Qwen2.5-VL 7B, Moondream2) + family order update in model_recommendation. Expanded curated list in model_capability_registry for better local VLM choices on Apple Silicon and other.
-- All additions checked for compatibility with existing KG (descriptions can be ingested), agents/tools dispatch, chat/computer_use (image pass-through), model rec logic, and non-multimodal fallbacks.
-
-- Perfect completion of report #15 recommended refactor:
-  - Server decomp wave: ModelRuntimeState class (not dict) for globals/wiring in model_runtime; sync_to_module_globals for compat; further clean.
-  - Deeper WorkspaceOSStore: timeline/plugins managers fully delegated and composed (record, has_permission etc all through).
-  - KG embed: set_embed_dim available for optional central handling.
-- All changes preserve legacy exactly, full composition, small modules.

package/docs/COMMUNITY_AND_PLUGINS.md

@@ -1,11 +1,11 @@
 # Community And Plugins
 
-Current release: **8.9.0 — Scoped Memory & Tool Policy Hardening**.
+Current release: **9.0.0 — Code Review Closure & Runtime Cleanup**.
 
 LatticeAI defines the path from a strong local-first framework (8.4.0
 action-aware baseline, 8.5.0 registry+DI hardening, 8.6.0 capture/navigation
 reliability, 8.7.0 runtime-state hygiene, 8.8.0 Brain Core extraction
-readiness, and 8.9.0 scoped Tool API hardening) to a product ecosystem. The
+readiness, 8.9.0 scoped Tool API hardening, and 9.0.0 cleanup closure) to a product ecosystem. The
 immediate goal is small and practical: make it clear how
 contributors can extend the Brain without weakening local-first trust,
 workspace scoping, or release quality.

package/docs/DEVELOPMENT.md

@@ -1,10 +1,10 @@
 # Lattice AI Development
 
-Current release: **8.9.0 — Scoped Memory & Tool Policy Hardening**.
+Current release: **9.0.0 — Code Review Closure & Runtime Cleanup**.
 
 This document is for contributors working on the local-first Digital Brain
 codebase. Product positioning and quick start stay in `README.md`; release
-history is intentionally limited to 7.0.0-8.9.0 in `docs/CHANGELOG.md` and
+history is intentionally limited to 8.0.0-9.0.0 in `docs/CHANGELOG.md` and
 `RELEASE.md`.
 
 ## Product Contract
@@ -114,10 +114,10 @@ For user-facing, API, runtime, release, or packaging changes, check:
 Release/publish examples must use exact target-version filenames. Do not
 document wildcard artifact upload commands.
 
-For 8.9.0 release work, exact artifacts are:
+For 9.0.0 release work, exact artifacts are:
 
-- `dist/ltcai-8.9.0-py3-none-any.whl`
-- `dist/ltcai-8.9.0.tar.gz`
-- `ltcai-8.9.0.tgz`
-- `dist/ltcai-8.9.0.vsix`
-- `src-tauri/target/release/bundle/dmg/Lattice AI_8.9.0_aarch64.dmg`
+- `dist/ltcai-9.0.0-py3-none-any.whl`
+- `dist/ltcai-9.0.0.tar.gz`
+- `ltcai-9.0.0.tgz`
+- `dist/ltcai-9.0.0.vsix`
+- `src-tauri/target/release/bundle/dmg/Lattice AI_9.0.0_aarch64.dmg`

package/docs/LEGACY_COMPATIBILITY.md

@@ -1,6 +1,6 @@
 # Legacy Compatibility Map
 
-Current target: **8.9.0 — Scoped Memory & Tool Policy Hardening**.
+Current target: **9.0.0 — Code Review Closure & Runtime Cleanup**.
 
 Lattice AI is moving toward a smaller, modular architecture centered on
 `lattice_brain`, `latticeai.services`, `latticeai.api`, and `latticeai.runtime`.

package/docs/ONBOARDING.md

@@ -1,6 +1,6 @@
 # Lattice AI Onboarding
 
-Current release: **8.9.0 — Scoped Memory & Tool Policy Hardening**.
+Current release: **9.0.0 — Code Review Closure & Runtime Cleanup**.
 
 The first-run goal is a five-minute path from "I opened the app" to "my Brain
 has a source, a question, and proof." This page is the product contract behind
@@ -32,7 +32,7 @@ read the docs first.
 
 ## Release Gate
 
-8.9.0 treats onboarding as a release gate, not marketing copy. The current
+9.0.0 treats onboarding as a release gate, not marketing copy. The current
 machine-checkable product readiness report requires this five-minute contract,
 the Brain Home surface, setup helpers, graph ingestion tests, and exact release
 artifact documentation before the release can be called complete.