feat(plane-project-os): integrate Plane multi-layout views, sprint cycles engine, triage intake, pure monochrome SVGs, and LangGraph supervisor multi-agent runtime

- Implemented Plane Multi-Layout Views: Kanban Board (4 columns), Dense List View (Linear style), Spreadsheet Grid View with inline quick edits, and 14-day Timeline Gantt view with 7 agent swimlanes.
- Built Plane Cycles Engine: sprint management (Cycle 102/103), dynamic burn-down calculations, story points tracking, and automated settlement on Founder Gate approval.
- Built Plane Triage Intake Gate: 1-click triage accept, reject with reason, and technical station assignment for founder briefs and swarm subtasks.
- Purged all legacy unicode glyphs and color noise across OrgChartView, replacing with 6 bespoke technical vector SVGs and pure monochrome tokens.
- Integrated LangGraph Supervisor Multi-Agent Engine: 5 specialized sub-graphs (PM Spec, Architect, Coder Sandbox, QA Security, Deliverables) with native HITL interrupt/resume and compaction checkpoints.
- Generated and verified 4 enterprise deliverables (Excel XLSX with SUMIFS, Word DOCX A4 spec, PowerPoint PPTX 16:9, PostgreSQL ACID migration SQL).
- Verified with 12/12 Playwright Plane E2E tests, 20/20 OrgChart tests, 17/17 Bun unit tests, and 0 TypeScript compilation errors.

feat(memory-service): implement Paperclip Document Revisions, Diff Viewer, and Collaborative Annotation Threads

- Added database models and repositories for DocumentRevision, DocumentAnnotationThread, DocumentAnnotationComment.
- Added REST APIs for document revisions (/api/documents/{id}/revisions) and annotations (/api/documents/{id}/annotations, /api/annotations/{id}/resolve).
- Built GitHub-style Split-View Diff Viewer (LCS dynamic line diff algorithm) with Transitions.dev spring motion.
- Added inline text annotation bubbles and multi-agent Slide-over Discussion Drawer in Memory Hub.
- Integrated automated cross-agent review workflow (Architect PRD critique -> PM resolution -> Coder/QA notification).
- Verified with 810/810 passing tests (544 Pytest + 266 Bun frontend/E2E) and clean Vite build.

feat(ai-company): complete AI Company OS with Company HQ Dashboard, Dynamic Agent Hiring, and AI Company Auditor Matrix

- Implemented 5-layer Python FastAPI backend with SQLite WAL mode and real-time SSE streaming.
- Built Pure Dark Mode Monochrome (#09090b) Company HQ Dashboard with Transitions.dev spring motion.
- Added Live Org Chart Tree, Deliverables Kanban with Founder Gate, Scoped Memory Hub, and Token Budget tracker.
- Built AI Company Auditor with 5-dimension scoring matrix (Grade A+ / 96 pts) and Scorecard Badge.
- Integrated Dynamic Agent Hiring Flow & On-Demand Org Growth based on Paperclip Core.
- Added 722+ comprehensive automated test cases with 100% pass rate.

feat(3d-banner): integrate ThreeUI Three.js 3D WebGL Neural Core banner and DeepSeek-V4 model family

feat(team): add AI workforce team management hub, gradient circle avatars, and collaboration flow DAG

feat(mock-fe): add multi-agent debate simulation, shadcn dark-light theme toggle, and standard lucide icons

All 6 completion criteria confirmed:
- grep canifa (source .py) = 0
- blueprint_id wired into POST /pipelines/run
- reports_to validator active
- 2-company DB constraint OK
- 5/5 blueprint E2E tests pass
- frontend build 0 errors
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Transforms Canifa-specific codebase into a generic AI Company platform
that can instantiate any business type from a YAML Company Blueprint.

Core additions:
- backend/schemas/blueprint.py: CompanyBlueprint, BlueprintRole,
  BlueprintPipelineNode, BlueprintSummary Pydantic models with
  reports_to cross-validation via @model_validator(mode='after')
- backend/services/blueprint_service.py: load/list/validate blueprints,
  deploy_company_from_blueprint() with 2-pass agent seeding for FK safety
- backend/blueprints/: 4 YAML blueprints (fashion-retail, software-studio,
  ecommerce-brand, marketing-agency) with roles, pipeline DAGs, nocobase_collections
- backend/tests/test_blueprint_e2e.py: 5 E2E tests — all pass

API changes:
- GET /companies/blueprints → List[BlueprintSummary] (with role_count)
- GET /companies/blueprints/{name} → full blueprint detail
- POST /companies/from-blueprint → deploy company from blueprint
- POST /pipelines/run accepts blueprint_id to drive orchestrator DAG

Orchestrator:
- ProjectOrchestrator now accepts pipeline_spec: list[dict] | None
- topological_sort_stages() (Kahn's BFS) derives parallel stages from blueprint DAG
- Falls back to PIPELINE_STAGES if no spec provided

De-branded:
- Removed all "Canifa" strings from models/services/api/tasks source files
- run_logs.py demo data → generic "Demo Brand/Server/Landing Page"
- agents.py demo HTML → "Demo Company"
- seed_db.py → uses deploy_company_from_blueprint("fashion-retail")

NocoBase:
- NocoBaseConnector: params-based init + from_company() classmethod
- company_secrets: UniqueConstraint('company_id', 'key') instead of global unique

Frontend:
- OnboardingWizard: blueprint picker with org-chart preview + DAG layer render
- blueprints.ts API client
- Companies.tsx: blueprint badge on company cards
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

- services/agent_log_writer.py: Writes ExecutionRunLog rows via AIAgent
  callbacks (tool_start, tool_complete, step, status). Errors swallowed
  to never block agent execution.

- services/workspace_manager.py: Creates isolated workspace directories
  per agent run (workspaces/{company_id}/{run_id}/) with src/,
  .agent_output/, README.md structure.

- services/path_security.py: Path validation to prevent directory
  traversal, blocks dangerous paths (C:\Windows, /etc, .env, .git).

- api/routes/run_logs.py: Added company-scoped runs_router with:
  GET /companies/{id}/runs/{runId} — run detail with all logs
  GET /companies/{id}/runs/{runId}/stream — SSE real-time log stream

- api/routes/agents.py: run_agent_in_background now creates workspace,
  instantiates AgentLogWriter, passes callbacks to AIAgent, and cleans
  up workspace in finally block.

- tests/test_agent_flow.py: Updated to mock WorkspaceManager and
  AgentLogWriter, relaxed AIAgent assertion to check core kwargs only.

All 3 tests pass. Dashboard verified 0 JS errors via Playwright.

Frontend calls (g ?? []).slice(0,10) on activity data, which crashes
when API returns {total, offset, limit, items} instead of ActivityEvent[].
Changed response_model from PaginatedResponse to List[ActivityLogResponse].
Also fixed .mcp.json Playwright package name.

Verified via Playwright MCP: 0 console errors on dashboard.

- Fix run_conversation sync/async boundary in 5 files:
  - run_agent.py main(): use _run_conversation_sync
  - batch_runner.py: use _run_conversation_sync
  - background_review.py: use _run_conversation_sync
  - curator.py: use _run_conversation_sync
  - acp_adapter/server.py: use _run_conversation_sync
- Fix classic Python closure late-binding bug in get_tools.py
  (is_async captured by reference, not value)
- Add missing deps to requirements.txt: openai, langchain-core, pyyaml, fire
- Fix missing HTTPException import in environment_config.py
- Fix missing List import in org_chart_svg.py

fix(api): change list_company_skills to return flat List[CompanySkillResponse] to fix frontend crash

feat: redesign UI to premium shadcn/ui with Canifa warm red/orange accents and translate company terms to Vietnamese departments (Phòng ban)

chore: Integrate Paperclip with FastAPI and resolve API case discrepancies, custom endpoints, and route issues

feat: migrate python backend, fix relative imports, fix adapter capabilities crash, and update README.md