Unclassified

v0 Prompt: IntegrateWise Platform App Shell — Universal Workbench for 12 Departments

Context Summary

This prompt builds the Platform App Shell (apps/platform) — the privileged control plane that manages the entire IntegrateWise platform. It inherits from the universal layer (apps/one-access) and provides department-specific workbenches for 12 departments.

What Already Exists (Don't Rebuild)

  • Universal Layer (apps/one-access): One Auth + Connectors, Marketplace, Discovery, Capabilities, Spine APIs
  • Kernel (@integratewise/kernel): 49 tests passing — MCP, Spine, Loader, Normalizer, ADK, Orchestrator, Discovery, Agent-to-Agent, Tool Memory
  • Tenant Spine Schema (@integratewise/spine-schema): Business context, KPIs, entities, connectors, agents, mappings
  • Web App (apps/web): Customer-facing business app with AppShell, WorkspaceShellNew, 12 domain configs, content router
  • 5 Platform Agents: Twin Orchestrator, Schema Generator, Schema Orchestrator, Adaptive Auto-Tuner, Coworker, Triage

What This Prompt Builds

The Platform App Shell — a single unified shell that serves as:

  1. Twin Orchestrator UI — central coordination surface
  2. 5 Platform Agent Workbenches — Schema Generator, Schema Orchestrator, Adaptive Auto-Tuner, Coworker, Triage
  3. 12 Department Workbenches — inherited from platform but customized per department
  4. Platform Control Plane — Configuration, Activation, Discovery, Capabilities, Governance, Memory, Deployments, Observability
  5. Inheritance Model — Platform features (auth, connectors, spine, projections) inherited from universal layer; UI shells owned here

Design Direction

Use the existing IntegrateWise design identity:

Quiet enterprise — Paper and forest — Neutral light and dark surfaces — Forest green accent (#1a3a2a) — Dense but highly readable operational layouts — Strong information hierarchy — Minimal decorative gradients — Subtle borders and restrained shadows — Desktop-first, responsive down to tablet and mobile — Accessible keyboard navigation

Use existing design-system tokens where possible. The UI should feel like one connected product, not a collection of unrelated dashboards.

Product Structure

Create a universal application shell with these top-level routes:

/platform
/platform/twin                    → Twin Orchestrator coordination surface
/platform/agents                  → Platform Agents overview
/platform/agents/schema-generator → Schema Generator workbench
/platform/agents/schema-orchestrator → Schema Orchestrator workbench
/platform/agents/adaptive-auto-tuner → Adaptive Auto-Tuner workbench
/platform/agents/coworker         → Coworker workbench
/platform/agents/triage           → Triage workbench

/platform/spine                   → Tenant Spine overview
/platform/spine/schema            → Schema Studio (tenant + source + mappings)
/platform/spine/business-context  → Business Context Studio
/platform/spine/kpis              → KPI Studio
/platform/spine/knowledge         → Business Knowledge workspace
/platform/spine/mappings          → Schema Mappings
/platform/spine/lineage           → Lineage graph

/platform/configuration           → Configuration Manager
/platform/activation              → Activation Manager
/platform/discovery               → Discovery Engine
/platform/capabilities            → Capability Registry
/platform/capability-matrix       → Capability Matrix
/platform/connected-tool-matrix   → Connected Tool Matrix
/platform/tool-capability-matrix  → Tool Capability Matrix

/platform/memory                  → Memory & Continuity
/platform/deployments             → Agent & Connector Deployments
/platform/governance              → Governance & Approvals
/platform/approvals               → Approval Workbench
/platform/observability           → Observability (agent comms, tool execs, data sharing)
/platform/agent-communication     → Agent Communication Log
/platform/tool-executions         → Tool Executions Log
/platform/data-sharing            → Data Sharing Audit

/platform/departments             → Department Registry (12 departments)
/platform/departments/[dept]      → Department-specific workbench (inherits platform shell)

/hosts                            → Host Registry
/hosts/[host]                     → Host Capability Projection

/grants                           → Grant Management
/activity                         → Unified Activity & Audit
/settings                         → Platform Settings

Global Application Shell

Create a complete shell with:

  • Product logo and product name ("IntegrateWise Platform")
  • Tenant selector (multi-tenant support)
  • Workspace selector (workbench context)
  • Host selector (IntegrateWise Web, IntegrateWise Platform, ChatGPT, Claude, MCP Client, Agent Runtime, External Application)
  • Environment selector (development, staging, production)
  • Global search (Cmd+K)
  • Command palette button
  • Notifications
  • Help
  • User profile menu
  • Collapsible left navigation (department-aware)
  • Context-sensitive right inspector
  • Breadcrumbs
  • Page title and page actions
  • Global status banner
  • Connection health indicator
  • Mock current user, tenant, workspace, and host context

Changing the host should visually update capability and access states using mock data.

1. Twin Orchestrator Surface (/platform/twin)

Purpose: Central coordination surface showing the Twin's intent interpretation, context resolution, agent selection, delegation, and result reconciliation.

Layout

Three-pane layout:

  • Left: Active Sessions / Execution Plans / Delegation Queue
  • Center: Current Execution Graph (live mermaid/flow visualization)
  • Right: Inspector — Context, Agent, Capabilities, Grants, Policies, Result

Key Components

Twin Status Card

┌─────────────────────────────────────────────────────────────┐
│  Twin Orchestrator          ● Active  ◐ Processing  ○ Idle  │
│  ─────────────────────────────────────────────────────────  │
│  Current Session: twin_sess_abc123                          │
│  Tenant: Acme Corp (tenant_001) | Workspace: CS-Prod        │
│  Host: IntegrateWise Platform | Env: production             │
│  ─────────────────────────────────────────────────────────  │
│  Intent: "Review at-risk accounts and prepare QBR"          │
│  Resolved Context: 12 accounts, 3 processes, 5 KPIs         │
│  Selected Agent: Account Health Agent (CS)                  │
│  Capabilities: CRM.read, Account.read, HealthScore.compute  │
│  Grants: grant_xyz (Salesforce prod, expires 2h)            │
│  Policy: Approval required for write actions                │
│  ─────────────────────────────────────────────────────────  │
│  [Pause] [Retry] [Escalate] [View Trace]                    │
└─────────────────────────────────────────────────────────────┘

Execution Graph (Center Pane)

  • Live directed graph showing: Intent → Triage → Twin → Agent(s) → Capabilities → Execution → Reconciliation
  • Node types: Intent, Context, Agent, Capability, Grant, Execution, Result, Memory Write, KPI Update
  • Edge types: delegates, resolves, invokes, grants, reconciles, updates
  • Click node → inspector shows details
  • Zoom, pan, fit view, minimap, time-window selector

Delegation Queue (Left Pane)

  • Table of pending/active/completed delegations
  • Columns: Agent, Function, Status, Entity Refs, Capabilities, Started, Duration, Result
  • Status badges: planned, waiting_context, waiting_approval, running, completed, failed, cancelled
  • Filter by agent, status, priority, time window

Inspector (Right Pane)

Tabs: Context | Agent | Capabilities | Grants | Policies | Result | Memory | Trace

2. Platform Agent Workbenches (5 Agents)

Each agent workbench follows a consistent pattern:

Shared Shell Components

Agent Header:
┌─────────────────────────────────────────────────────────────┐
│  [Icon] Schema Generator          Status: ● Active          │
│  Category: Platform Intelligence | Function: Schema Generation│
│  Tenant: Acme Corp | Workspace: Platform | Host: Platform   │
│  ─────────────────────────────────────────────────────────  │
│  Business Processes: [schema_discovery, schema_mapping]     │
│  Objectives: [data_integrity, semantic_consistency]         │
│  KPIs: [mapping_confidence, schema_coverage, breaking_changes]│
│  Entities: [source_schema, tenant_entity, field_mapping]    │
│  Knowledge: [naming_conventions, type_mapping_rules]        │
│  Capabilities: [discovery.read, schema.propose, mapping.write]│
│  Connections: [PostgreSQL, MySQL, OpenAPI, GraphQL, MCP]    │
│  Policies: Risk: medium | Approval: schema_activation       │
│  Runtime: model: gpt-4o | deployment: platform-agent-v1     │
│  ─────────────────────────────────────────────────────────  │
│  [Run] [Configure] [Evaluate] [Deploy] [Audit]              │
└─────────────────────────────────────────────────────────────┘

2.1 Schema Generator (/platform/agents/schema-generator)

Left Pane: Discovery Runs

  • Table: Run ID, Source, Status, Entities Detected, Fields, Relationships, Confidence, Created, Actions
  • Actions: View Proposals, Re-run, Configure Sources

Center Pane: Proposed Schema Canvas

  • Visual entity-relationship diagram of proposed tenant schema
  • Nodes: Entity Types (proposed), Fields, Relationships
  • Edge: has_field, relates_to, maps_from
  • Color coding: high confidence (green), medium (amber), low (red), conflict (purple)
  • Click entity → field detail panel

Right Pane: Proposal Inspector
Tabs: Overview | Entities | Fields | Relationships | Mappings | Conflicts | Confidence | Lineage

2.2 Schema Orchestrator (/platform/agents/schema-orchestrator)

Left Pane: Schema Lifecycle

  • Kanban board: Discovered → Generated → Mapped → Contextualized → Validated → Reviewed → Approved → Activated → Monitored
  • Cards show: Entity count, Breaking changes, Reviewers, Approval status

Center Pane: Active Schema Version

  • Current tenant schema visualization with version history
  • Diff view between versions
  • Breaking change detector

Right Pane: Orchestration Controls
Tabs: Discovery | Mapping Review | Business Context | KPI Binding | Capability Binding | Validation | Approval | Activation | Rollback

2.3 Adaptive Auto-Tuner (/platform/agents/adaptive-auto-tuner)

Left Pane: Tuning Recommendations

  • Table: Area, Current Value, Recommended, Expected Benefit, Risk, Affected Agents, Approval Required, Status
  • Status: observe, recommend, simulate, canary, approve, apply, monitor, rollback

Center Pane: Impact Simulation

  • Before/after charts for latency, cost, failure rate, approval rejection rate, KPI trends
  • Canary deployment configuration

Right Pane: Tuning Detail
Tabs: Observations | Recommendations | Simulations | Canary Results | Approvals | Applied | Monitoring | Rollback Config

2.4 Coworker (/platform/agents/coworker)

Left Pane: Active Collaborations

  • List of user sessions the Coworker is assisting
  • Columns: User, Role, Department, Workbench, Current Task, Intent, Status

Center Pane: Coworker Workspace

  • Shows the user's current workbench with Coworker overlay
  • Structured intent builder: Category → Function → Entities → Process → KPIs → Capabilities
  • "Send to Twin" button

Right Pane: Coworker Context
Tabs: User Profile | Workbench State | Available Capabilities | Recent Memory | Pending Approvals | Corrections

2.5 Triage (/platform/agents/triage)

Left Pane: Incoming Work Queue

  • Table: Source, Category, Function, Priority, Risk, Entities, Process, KPIs, Recommended Agent, Capabilities, Approval Required, Missing Context, Duplicate Of, Created
  • Sources: User, Email, Ticket, Alert, Incident, Connector Failure, Agent Failure, Schema Conflict, Approval Request, Deployment Event, Discovery Event

Center Pane: Triage Detail

  • Selected item detail with classification confidence
  • Entity enrichment suggestions
  • Process/KPI matching

Right Pane: Routing Decision
Tabs: Classification | Enrichment | Routing | Escalation | Audit

3. Department Workbenches (12 Departments)

Each department inherits the Platform Shell but customizes:

Department Configuration (from workspace-config.ts)

// 12 Departments:
(CUSTOMER_SUCCESS,
  SALES,
  REVOPS,
  MARKETING,
  PRODUCT_ENGINEERING,
  FINANCE,
  SERVICE,
  PROCUREMENT,
  IT_ADMIN,
  STUDENT_TEACHER,
  PERSONAL,
  BIZOPS);

Department Workbench Route: /platform/departments/[dept]

Inherits from Platform:

  • Global shell (tenant, workspace, host, env selectors)
  • Twin Orchestrator access (Cmd+K or header button)
  • Platform agents (via navigation)
  • Spine, Configuration, Activation, Discovery, Capabilities
  • Governance, Memory, Deployments, Observability
  • One Auth + Connectors (auth, marketplace, connectors, grants)

Department-Specific:

  • Domain theme (accent color, icon, label)
  • Work navigation sections (L1, Intelligence, Tools, etc.)
  • Dashboard metrics (5 priority metrics)
  • Primary entity + secondary entities
  • Connector categories
  • Entity types for workbench
  • Happy path modules (Today, Queue, Decisions)

Department Workbench Layout

┌─────────────────────────────────────────────────────────────────────────────┐
│ IntegrateWise Platform  [Tenant ▼] [Workspace ▼] [Host ▼] [Env ▼] [Search]  │
│ ──────────────────────────────────────────────────────────────────────────  │
│ [Platform] [Departments ▼] [Spine] [Config] [Agents] [Gov] [Obs] [Settings] │
├──────────────┬──────────────────────────────────────────────────┬───────────┤
│              │                                                  │           │
│  NAVIGATION  │              DEPARTMENT WORKBENCH                │ INSPECTOR │
│              │                                                  │           │
│  [L1]        │  ┌──────────────────────────────────────────┐   │  Context  │
│  ├ Dashboard │  │  Department Header                       │   │  Entity   │
│  ├ Accounts  │  │  Title: "Account Success"                │   │  KPIs     │
│  ├ Today     │  │  Industry: SaaS · CS · CSM               │   │  Caps     │
│  ├ Tasks     │  │  Health: ● 92%                           │   │  Grants   │
│  ├ Meetings  │  └──────────────────────────────────────────┘   │  Policies │
│  ├ Renewals  │  ┌─────────┬─────────┬─────────┬─────────┐    │  Memory   │
│  ├ Risk Mtrx │  │  ARR    │ Health  │ At-Risk │ Expand  │    │  Trace    │
│  ├ At-Risk   │  │ $2.4M   │ 87%     │ 12      │ $480K   │    │           │
│  ├ Health    │  └─────────┴─────────┴─────────┴─────────┘    │           │
│  ├ API Port. │  ┌─────────────────┬───────────────────────┐   │           │
│  ├ Engage.   │  │ Active Entities │ Governance Queue      │   │           │
│  [Intel]     │  │ ┌─────────────┐ │ ┌─────────────────┐  │   │           │
│  ├ BI        │  │ │ Acme Corp   │ │ │ Schema Change   │  │   │           │
│  ├ Insights  │  │ │ Health: 92% │ │ │ Review Required │  │   │           │
│  ├ Queue     │  │ │ ARR: $2.4M  │ │ │ [Approve][Reject]│  │   │           │
│  ├ Decisions │  │ └─────────────┘ │ └─────────────────┘  │   │           │
│  [Tools]     │  │ ... (6 entities)  ... (3 items)       │   │           │
│  ├ Knowledge │  └─────────────────┴───────────────────────┘   │           │
│  ├ Workflows │                                                  │           │
│  ├ Integ.    │                                                  │           │
│  ├ Settings  │                                                  │           │
│              │                                                  │           │
│  [Personal]  │                                                  │           │
│  ├ Calendar  │                                                  │           │
│  ├ Notes     │                                                  │           │
│  ├ Hub Home  │                                                  │           │
│  ├ Knowledge │                                                  │           │
│              │                                                  │           │
└──────────────┴──────────────────────────────────────────────────┴───────────┘

Department Workbench Features

1. Metric Tiles (5 per department)

  • Real-time values from Spine projection
  • Types: currency, number, percentage, trend
  • Click → drill into entity list

2. Operational Canvas (Two-column)

  • Left: Active Entities (cards with health score, status, sources, actions)
  • Right: Governance Queue (approve/reject for elevated roles)

3. Three Tabs (Work | Organization | Personal)

  • Work: Department-specific operational view
  • Organization: Shared context (OrgView component)
  • Personal: Personal productivity (PersonalView component)

4. Twin Intelligence Slot

  • Contextual Twin suggestions on shell modules (dashboard, today, workbench)
  • Max 1 item, dismissible

5. Content Router Integration

  • Non-shell modules render via AnimatedContentRouter
  • Lazy-loaded per domain module registry
  • Density-gated empty states

4. Platform Control Plane Surfaces

4.1 Spine Overview (/platform/spine)

Graph Canvas — Semantic and operational network view

Node Types (16):
Tenant, Workspace, Host, User, Agent, Connector, Tool, Capability, Entity Type, Entity, Business Process, Business Objective, KPI, Knowledge Source, Memory, Policy, Execution, Deployment

Edge Types (15):
Provides, Uses, Reads, Writes, Triggers, Calls, Shares data with, Authorized by, Governed by, Affects KPI, Participates in process, Mapped from, Deployed to, Depends on

Controls:
Zoom, Pan, Fit view, Minimap, Filters, Grouping, Expand neighbors, Collapse groups, Time-window selector, Live-event toggle, Layout selector, Search, Path tracing, Upstream/downstream tracing

Node Inspector: Metadata, Business meaning, Related capabilities, Related KPIs, Related hosts, Related agents, Source mappings, Activity, Health, Policies, Lineage

4.2 Schema Studio (/platform/spine/schema)

Three-pane layout:

  • Schema Navigator (tree): Generic Spine → Tenant Schema → Source Schemas → Entity Types → Fields → Relationships → Events → Domain Packs → Versions
  • Schema Canvas/Table (center): Visual ERD or field table for selected entity type
  • Schema Inspector (right): Technical name, Business name, Description, Fields, Relationships, Source mappings, Business contexts, KPIs, Capabilities (read/write), Agents, Policies, Lineage

4.3 Business Context Studio (/platform/spine/business-context)

Three-pane layout:

  • Context Tree (left): Organization → Industry → Departments → Functions → Objectives → Processes → Terminology → Policies → Roles → Owners → Related KPIs → Related Entities → Related Capabilities → Related Knowledge
  • Context Canvas (center): Visual relationship map
  • Context Inspector (right): Full detail with create/link/version actions

Supports: Create context, Link entity, Link KPI, Link capability, Link process, Link knowledge, Version history, Compare versions, Publish draft

Mock examples: Software development, Sales, Finance, Operations — but UI is generic

4.4 KPI Studio (/platform/spine/kpis)

Views: KPI Catalog | Definitions | Targets | Dimensions | Data Sources | Related Processes | Related Capabilities | Owners | Calculation History

Visual differentiation:

  • Raw metric (neutral)
  • Derived metric (blue)
  • Business KPI (green)
  • Operational indicator (amber)
  • Agent metric (purple)
  • Platform health metric (red)

KPI Detail: Name, Business purpose, Formula, Unit, Current value, Target, Warning threshold, Critical threshold, Trend, Data sources, Related entities, Related processes, Related capabilities, Owners, Agent influence, Version, Last calculated

4.5 Business Knowledge (/platform/spine/knowledge)

Knowledge Types: Policy, SOP, Runbook, Playbook, Contract, Product documentation, Process documentation, Decision, Meeting outcome, Exception, Lesson learned

Workspace: Source list | Document viewer | Provenance | Related business context | Related entities | Related capabilities | Related KPIs | Related agents | Freshness | Sensitivity | Retention | Version

4.6 Schema Mappings (/platform/spine/mappings)

Source-to-tenant mappings table:
Source system | Source object | Tenant entity type | Field mappings | Transformations | Confidence | Mapping status | Reviewer | Version | Last sync | Conflicts | Validation errors

Visual Mapper: Drag-and-drop mapping lines, Suggested mappings, Conflict resolution, Mapping inspector, Draft → Review → Approve → Activate states

4.7 Lineage (/platform/spine/lineage)

Graph mode | Timeline mode | Table mode
Upstream trace | Downstream trace | Sensitive-data filter | Export | Impact analysis

Columns: Source system → Source object → Transformation → Tenant entity → Capability → Agent → Workbench → KPI → Destination → Memory write

5. Integration Control Plane

5.1 Configuration Manager (/platform/configuration)

Sections (tabs): Providers | Tenant | Workbench | Twin | Memory | Notifications | Retention

Each section: Key-value editor with validation, schema-driven forms, environment overrides, version history, rollback

5.2 Activation Manager (/platform/activation)

Pipeline view: Discovered → Configured → Validating → Activating → Active → Degraded → Suspended

Per connection: Provider, Connection name, Tenant, Connected account, Auth method, Granted scopes, Authorized hosts, Used by agents, Health, Last sync, Status, Expiry, Actions

Detail tabs: Overview | Authorization | Configuration | Capabilities | Source Schema | Tenant Schema | Schema Mappings | Business Context | KPI Impact | Activation | Executions | Health | Lineage | Audit

Flows: Mock reconnect, Scope review, Host assignment

5.3 Discovery Engine (/platform/discovery)

Three-pane layout:

  • Discovery Sources (left): Host platforms, Workspace connections, Connector catalog, MCP servers, OpenAPI specs, GraphQL schemas, Databases, Plugins, Agents, Knowledge sources
  • Discovered Resources (center): Table with type badges, confidence, status, proposed mapping
  • Discovery Inspector (right): Technical metadata, Auth requirements, Input/output schemas, Source authority, Duplicate matches, Proposed tenant mapping, Business-context relationships, KPI relationships, Risk, Approval state

Controls: Start discovery, Progress, Live events, Source health, Deduplication status

5.4 Capability Registry (/platform/capabilities)

Sections: Available to me | Installed | Connector capabilities | MCP tools | Plugin capabilities | Agent capabilities | AI capabilities | Native platform capabilities | Recently used | Recommended

Capability Card: Name, Provider, Type, Read/Write, Risk, Approval requirement, Connected state, Host availability, Agent availability, Input schema summary, Output schema summary, Health, Last used

Detail tabs: Overview | Input schema | Output schema | Permissions | Required scopes | Hosts | Agents | Connections | Business context | KPI impact | Execution history | Policies | Versions | Audit

Actions: Run | Add to workbench | Assign to agent | Grant to host | Add to automation | Ask Twin | Inspect schema | View lineage

Schema-driven invocation form using mock JSON schema

6. Matrix Views

6.1 Capability Matrix (/platform/capability-matrix)

Rows: Connectors | Tools | Agents | Hosts | Roles | Workspaces
Columns: Capabilities | Permissions | Environments | Providers | Host platforms

Cell States (10): Available | Connected | Missing auth | Missing permission | Approval required | Policy denied | Provider degraded | Not installed | Unsupported by host | Configuration required

Features: Freeze rows/cols, Search, Filters, Grouping, Column selection, Density toggle, Export, Compare mode, Cell inspector, Mock bulk assignment

6.2 Connected Tool Matrix (/platform/connected-tool-matrix)

Rows (switchable): Hosts | Applications | Agents | Workspaces | Tenants
Columns: Slack, GitHub, Stripe, Salesforce, Notion, Snowflake, OpenAI, Anthropic, Custom MCP Server, etc.

Cell States: Connected | Shared | Authorized | Expired | Revoked | Configuration required | Policy denied | Available but unassigned

Cell Inspector: Connection, Granted scopes, Authorized capabilities, Assigned host/agent, Environment, Last use, Grant expiry, Policy, Audit events

6.3 Tool Capability Matrix (/platform/tool-capability-matrix)

Rows: Tools/Providers (nested groups)
Columns: Read | Create | Update | Delete | Search | Trigger | Subscribe | Execute | Export | Administrative

Example GitHub:

  • Repository: read ✓, create ✓, update ✓, delete ✗, search ✓
  • Issue: read ✓, create ✓, update ✓, delete ✗
  • Pull Request: read ✓, create ✓, update ✓, merge ✓
  • Workflow: trigger ✓

Cell States: Supported | Unsupported | Scope missing | Approval required | Restricted | Host incompatible | Agent restricted

7. Memory & Continuity (/platform/memory)

Tabs: Working Memory | Task Memory | Entity Memory | Agent Memory | Operational Memory | Historical Decisions | Lessons Learned

Memory Entry: ID, Type, Entity Ref, Agent Ref, Content, Embeddings, Provenance, Created, Updated, Access Count, Sensitivity, Retention Policy, Version

Actions: Search, Filter by type/agent/entity/time, Export, Delete (with audit), Compare versions

8. Deployments (/platform/deployments)

Tabs: Agent Deployments | Connector Deployments | Model Deployments | Environment Status

Agent Deployment: Agent ID, Category, Function, Version, Model, Environment, Tenant, Status, Health, Last Deployed, Rollback Version, Config

Actions: Deploy, Rollback, Scale, Configure, View Logs, Evaluate

9. Governance & Approvals

9.1 Governance (/platform/governance)

Tabs: Permissions | Policies | Approval Rules | Data Sharing Controls | Risk Classification | Audit

Policy Editor: Schema-driven policy definition with conditions, actions, targets, exceptions

9.2 Approvals (/platform/approvals)

Unified queue: Schema changes, Tool invocations, Data sharing, Agent deployments, Configuration changes, Access grants

Approval Item: Requestor, Type, Target, Risk, Context, Required Approvers, Current Votes, Deadline, Actions: Approve, Reject, Defer, Request Info

10. Observability

10.1 Observability Overview (/platform/observability)

Dashboard: Agent Communication | Tool Executions | Data Movement | KPI Influence | Failures | Costs | Traces

10.2 Agent Communication (/platform/agent-communication)

Log table: Timestamp, Source Agent, Target Agent, Twin Session, Intent, Capabilities, Status, Duration, Trace ID

Detail: Full message payload, Context, Policy checks, Result

10.3 Tool Executions (/platform/tool-executions)

Log table: Timestamp, Actor (User/Agent/Automation), Tool, Capability, Input (truncated), Output (truncated), Status, Duration, Cost, Trace ID, Tenant, Workspace, Host

Filters: By actor, tool, capability, status, time, tenant, risk

10.4 Data Sharing (/platform/data-sharing)

Log table: Timestamp, Source, Destination, Entity Types, Fields, Purpose, Grant ID, Policy, Consent, Risk

11. Hosts & Grants

11.1 Hosts (/hosts)

List: IntegrateWise Web, IntegrateWise Platform, ChatGPT, Claude, MCP Clients, Agent Runtimes, Plugins, External Applications

Host Detail: Host type, Client ID, Status, Allowed scopes, Granted connections, Available capabilities, Unsupported capabilities, Active sessions, Grants, Last activity, Environment, Policies, Audit

Capability Projection Preview: What this host would receive (MCP tool list, scopes, entity access)

11.2 Grants (/grants)

Table: Grant ID, Subject, Audience, Tenant, Workspace, Host, Connection, Capability scopes, Purpose, Environment, Issued, Expires, Status, Policy version, Last used

Actions: Revoke, Extend, Duplicate, Inspect

Filter: Expiring soon, By host, By agent, By capability

Grant Creation Drawer: Mock state

12. Activity & Audit (/activity)

Unified Event Types: Login, Consent, Connection created, Connection revoked, Discovery started, Capability discovered, Schema mapped, Grant issued, Tool invoked, Approval requested, Policy denied, Data shared, KPI updated, Host projection generated

Views: Timeline | Table
Filters: Actor, Tenant, Host, Agent, Tool, Outcome, Risk, Correlation ID
Trace Inspector: Full event detail with context

Shared Reusable Components

Create these as reusable components (used across all surfaces):

  • Status badges (healthy, degraded, expired, pending, denied)
  • Risk badges (low, medium, high, critical)
  • Host badges (with icons)
  • Provider logos
  • Capability cards
  • Connection cards
  • Matrix cells (with state icons, tooltips)
  • Schema viewers (ERD, table, JSON)
  • JSON schema forms (dynamic, validation)
  • Timeline events
  • Graph node inspector
  • Filter toolbar (multi-select, search, save)
  • Search command (Cmd+K)
  • Empty states (configured, discovered, connected, error)
  • Loading states (skeleton, spinner, progress)
  • Error states (retry, dismiss, support)
  • Unauthorized state
  • Forbidden state
  • Offline state
  • Stale-data banner
  • Approval banner
  • Mock notifications

Mock Data Requirements

Provide rich mock datasets for:

  • Multiple tenants (Acme Corp, Globex, Wayne Enterprises, Stark Industries)
  • Multiple workspaces per tenant (Production, Staging, Development)
  • Multiple hosts (all 7 host types)
  • At least 12 connectors (Salesforce, HubSpot, Slack, GitHub, Stripe, Snowflake, Notion, Jira, OpenAI, Anthropic, Custom MCP, PostgreSQL)
  • At least 30 capabilities (CRUD, search, trigger, subscribe, execute, export, admin per provider)
  • At least 8 agents (5 platform + 3 category examples)
  • At least 15 business entities (Account, Contact, Deal, Campaign, Invoice, Ticket, Incident, Feature, Sprint, Release, etc.)
  • At least 10 KPIs (ARR, Health, Pipeline, Conversion, Uptime, MTTR, NPS, Churn, CAC, LTV)
  • At least 10 knowledge sources (Policy, SOP, Runbook, Contract, Decision, Lesson)
  • At least 50 activity events across all types
  • Multiple schema mappings (source → tenant)
  • Multiple grant states (active, expiring, expired, revoked, pending)
  • Multiple connection states (connected, degraded, expired, revoked, configuration required)

The UI should demonstrate:

  • Healthy states
  • Empty states
  • Loading states
  • Partial data
  • Degraded providers
  • Expired authorization
  • Policy denial
  • Approval required
  • Unsupported host capability
  • Schema conflict
  • Stale business context

Technical Constraints

Frontend Only

  • React + TypeScript
  • Next.js App Router conventions
  • Tailwind CSS
  • Reusable components
  • Local mock data
  • Local component state

No Backend Implementation

  • No real authentication
  • No real OAuth
  • No real API calls
  • No database
  • No secrets
  • No external provider credentials
  • No actual token creation
  • No real MCP connection
  • No provider SDKs
  • No hardcoded business-domain assumptions in reusable components

Create production-quality UI shells with:

  • Realistic data density
  • Navigation
  • Detail pages
  • Inspectors
  • Tables
  • Matrices
  • Graphs
  • Drawers
  • Dialogs
  • Filters
  • Responsive states

Inheritance Model (Critical)

apps/platform (Platform App Shell)
├── Inherits from apps/one-access (Universal Layer):
│   ├── /auth → One Auth (identity, tenant, consent, grants, host auth)
│   ├── /marketplace → Connector catalog, marketplace
│   ├── /connectors → Connection management, health, config
│   ├── /discovery → Discovery engine
│   ├── /capabilities → Capability registry
│   ├── /connected-tool-matrix → Who connects to what
│   ├── /tool-capability-matrix → What tools can do
│   └── /spine → Tenant Spine, Business Context, KPIs, Knowledge, Mappings, Lineage
│
├── Owns (Platform UI Shells):
│   ├── Twin Orchestrator surface
│   ├── 5 Platform Agent workbenches
│   ├── 12 Department workbenches (customized shells)
│   ├── Configuration Manager
│   ├── Activation Manager
│   ├── Discovery Engine UI
│   ├── Capability Matrix UIs
│   ├── Memory & Continuity
│   ├── Deployments
│   ├── Governance & Approvals
│   ├── Observability (Agent Comms, Tool Execs, Data Sharing)
│   ├── Host Registry & Projection Preview
│   ├── Grant Management
│   └── Activity & Audit
│
└── apps/web (Business App) CONSUMES Platform:
    ├── Department workbenches (via platform shells)
    ├── Marketplace projections
    ├── Coworker experience
    └── Business-user Twin interface

Department-to-Platform Mapping

DepartmentDomain Config KeyPrimary EntityAccent ColorKey Modules
Customer SuccessCUSTOMER_SUCCESSAccount#1a3a2aDashboard, Accounts, Today, Tasks, Meetings, Renewals, Risk Matrix, At-Risk, Health, API Portfolio, Engagement, BI, Insights, Queue, Decisions, Knowledge, Workflows, Integrations, Settings
SalesSALESDeal#2d5a3dDashboard, Pipeline, Pipeline Kanban, Deals, Contacts, Activities, Forecasting, Quotes, Sequences, Sales Analytics, Sales Today, Sales Queue, Sales Decisions
RevOpsREVOPSRevenue#28332cDashboard, Pipeline, Forecasting, Quotas, Analytics, Cohorts, Team, Metrics, Pending Review, Decision Log, RevOps Today, RevOps Queue, RevOps Decisions
MarketingMARKETINGCampaign#b8943fDashboard, Campaigns, Leads, Analytics, Email, Social, Attribution, Forms, Blog, Website, SEO, Media, Pages, Theme, Marketing Today, Marketing Queue, Marketing Decisions
Product & EngineeringPRODUCT_ENGINEERINGIncident#2d5a3dDashboard, Incident Queue, Roadmap, Features, Bugs, Sprints, Tasks, Releases, Analytics, Feedback, PE Today, PE Queue, PE Decisions, Fabric Admin, Settings
FinanceFINANCERevenue#2d5a3dDashboard, Revenue, Expenses, Invoices, Invoice Approvals, Reports, Forecasting, Budget, Data Governance, Finance Today, Finance Queue, Finance Decisions
ServiceSERVICETicket#1a3a2aDashboard, SLA Dashboard, Tickets, Customers, Knowledge, Analytics, Satisfaction, Service Today, Service Queue, Service Decisions, Fabric Admin
ProcurementPROCUREMENTContract#1a3a2aDashboard, Renewals, Vendors, Orders, Contracts, Spend, Savings, Procurement Today, Procurement Queue, Procurement Decisions
IT AdminIT_ADMINSystem#28332cDashboard, Systems, Incidents, Vulnerabilities, Certificates, Licenses, IT Admin Today, IT Admin Queue, IT Admin Decisions
Student TeacherSTUDENT_TEACHERStudent#b8943fDashboard, At-Risk, Students, Courses, Assignments, Gradebook, Student Today, Student Queue, Student Decisions
PersonalPERSONALTask#6b7280Dashboard, Hub Home, Hub Today, Founder Today, Founder Projects, Founder Decisions, Knowledge Hub, What's New, Tasks, Calendar, Notes, Projects, Docs, Bookmarks, Wealth Pilot, Wellness Core, Learning Desk, Relationship Map, Fabric Admin, Settings
BizOpsBIZOPSRevenue#2d5a3dDashboard, Ops Command Center, Strategic Hub, Metrics, CRM, Leads, Sales Hub, Clients, Projects, Workflows, Analytics, Accounts, Tasks, Docs, Calendar, Integrations, Workflow Canvas, BizOps Today, BizOps Queue, BizOps Decisions, Founder Ops, CEO Dashboard, COO Dashboard, CIO/CTO Dashboard, 14 Department Unified Views