Dashboard Architecture

Dashboard Architecture 문서는 OpenClaw 공식 문서(web/dashboard-architecture)를 한국어로 정리한 가이드입니다. Session dashboards: architecture and implementation plan (technical design, pre-GA) 명령·설정 키·코드 예시는 공식 문서를 그대로 보존하며, 해석과 절차 안내는 한국어로 제공합니다. 최종 동작은 설치된 CLI 버전과 공식 원문을 확인하세요.

핵심 요약

Session dashboards: architecture and implementation plan (technical design, pre-GA)

한국어 가이드 범위: web/dashboard-architecture 경로의 설정·명령·제약·예시를 학습용으로 재구성합니다.

문서 구성

공식 문서의 주요 섹션은 다음과 같습니다.

Vision
개념
UX flows
Interaction tiers
Widget model and hosting
Widgets host content; MCP apps are one content kind
Plugin capability declarations
Modeled residual: WebRTC data channels
Transcript display: one widget card
Server-sourced widgets (pinned MCP apps)
WorkBoard integration
Layout: fluid grid
Data model (per-agent DB)
Protocol surface
Agent tools
What this replaces
Non-goals (this program)
Implementation plan

상세 내용

본문

Technical design document for the session dashboard feature, written before and during implementation. It is the source of truth for the build-out. When the feature ships, /web/dashboard becomes the user-facing page and this page stays as the architecture reference.

Vision

Working with an agent today is a text stream. The dashboard makes it a workbench: the agent renders live, interactive widgets; the user pins them onto a persistent surface; chat docks to the side (or hides) and the main content is the board. You go from "talking to the agent" to "operating a control panel the agent built for you" without ever leaving the session.

주요 항목:

A board is a face of a session, not a new object. Every session (thread)
Agent parity. Everything the user can do on a board, the agent can do
Native, not embedded. The board is Lit components in the Control UI shell
Small agent surface. Widgets are addressed by stable name and updated in
Capabilities over trust. Widget code is arbitrary agent-authored HTML/JS

개념

Concept	Definition
Session (thread)	Existing gateway session, keyed by stable `sessionKey`. Owned by an agent.
Board	The widget face of one session. Exists iff the session has widgets/tabs. Survives `/new`/`/reset` (attached to `sessionKey`, not the transcript).
Tab	A presentation page of a board: which widgets, their arrangement, and the chat dock state (`left`/`right`/`bottom`/`hidden`). Boards start with one implicit tab.
Widget	Named, sandboxed HTML/JS program owned by the session. Addressed as `sessionKey` + `name`. Updated in place by name.
Capability manifest	Per-widget declaration of reach: `data` (read bindings), `actions` (allowlisted verbs), `prompt` (send to session), `net` (allowed origins).
Pin (widget)	Moving a transcript widget onto the session's board (user affordance or agent tool arg). Unpin removes it from the board.
Pin (session)	Existing sidebar pinning of sessions. A pinned session with a board opens on its board face.

UX flows

in the transcript exactly as today → hover shows Pin to dashboard → widget appears on the session's board. The agent can pass pin: true to do the same.

주요 항목:

Graduation: agent calls show_widget in any chat → widget renders inline
Board view: a session with a board gets a face toggle (Chat / Dashboard).
Drag: user drags widgets; grid auto-compacts (widgets float up, neighbors
Reset warning: /new / /reset on a board-bearing session asks for
Sidebar: pinned sessions render their board face when they have one.
Interactions (three tiers, see below): silent state events, visible

Interaction tiers

State events (default). Widget UI interactions the model should know about but not respond to. bridge.emitState({...}) appends a structured session notice (same mechanism as group-activity notices). No agent turn is started; the model sees accumulated notices on its next run. 2. Prompts (explicit talk). bridge.sendPrompt(text) — requires user activation; sends a visible user message into the session (the docked chat shows it). Rate-limited; each send is user-confirmed unless the widget holds the prompt capability grant. 3. Automation. bridge.runAction(name, args) — fires a manifest-declared action. Initial verb set: cron.trigger (run an existing cron job now) and binding.refresh. Cron jobs already run in visible, isolated run-sessions and can use a cheaper model: that is the "small model powers the widget" path. No hidden sessions anywhere.

Widget HTML/JS is authored by the agent (typically via show_widget), wrapped in the standard document shell (CSP meta, size reporter, bridge bootstrap) and rendered in `` (never allow-same-origin).

주요 항목:

Inline (transcript) widgets keep the current canvas-document pipeline:
Board widgets are session state: bytes live in the owning agent's SQLite
Update in place: re-emitting a widget with the same name replaces the
Byte freezing: granted capabilities bind to the sha256 of the widget

Widgets host content; MCP apps are one content kind

The widget is the OpenClaw primitive: the named, pinned, sized, session-owned board cell with a grant record. What renders inside it is a content kind:

주요 항목:

html — agent-authored via show_widget, bytes in board storage.
mcp-app — a third-party MCP app view (ui:// resource from a configured
One sandbox host. html widgets render through the same hardened
One authorization model. A widget's reach is a granted allowlist,
Host tools for html widgets (exposed over the widget bridge, checked
openclaw.prompt.send — tier 2; routed through the visible composer,
openclaw.state.emit — tier 1 session notices (coalesced, size-capped)
openclaw.data.read — parameterized read-only bindings (existing
openclaw.cron.trigger — tier 3 automation
net = CSP. Network reach uses the already-shipped per-widget CSP
Grants. A widget declaring nothing renders immediately (sandboxed,
Authoring shim. The document wrapper injects window.openclaw.prompt,

Plugin capability declarations

Enabled plugins can extend the widget host through dashboard.dataBindings and dashboard.actionVerbs in openclaw.plugin.json. Plugin-local ids become grant names prefixed by the plugin id, such as workboard.cards.list and workboard.dispatch; % and . in the plugin-id segment are escaped so a different plugin/local-id split cannot inherit the same persisted grant. During plugin registration, OpenClaw verifies that every binding targets an RPC registered by the same plugin with operator.read and every action targets one with operator.write; invalid declarations fail the plugin load. The validated registry is rebuilt only with plugin lifecycle changes, while widget grants remain per-widget and byte-and-revision-bound.

Modeled residual: WebRTC data channels

The sandbox CSP emits the proposed webrtc 'block' directive, but Chromium's current CSP directive set does not implement it. Scriptable widgets can therefore use WebRTC data channels for egress in current Chromium. The same residual already ships for inline chat widgets and the MCP Apps host on main.

Inline display unifies on the widget primitive. When a tool result carries UI — show_widget output or an MCP tool result with an app resource — the system materializes an ephemeral, auto-named widget (session-scoped, pruned) and the transcript renders a single widget card that dispatches on content kind. MCP app auto-display stays exactly as the spec expects (zero extra model work); it just is a widget underneath. This deletes the parallel mcpApp special-cases in chat rendering (surface gating, separate dedup), gives every inline UI the same pin affordance, and makes the widget registry the primary re-open path (transcript-scan reconstruction stays as fallback for never-pinned history). The read-only ticketed standalone host overlaps with boards as a persistent re-open surface — consolidation candidate to evaluate in T6, not assumed.

Server-sourced widgets (pinned MCP apps)

With the unified host, pinning a third-party MCP app is just a widget whose content is fetched from the server instead of stored: board_widgets keeps the descriptor (serverName, toolName, uiResourceUri, originating toolCallId + sessionKey) instead of HTML bytes, and the board re-mints the view lease past the chat-turn 10-minute TTL (re-fetching the ui:// resource on staleness). Chat inline MCP app views get the same Pin to dashboard affordance as agent widgets. Re-opened views are read-only today by design; pinned apps that should stay interactive get a durable grant over the server's app-visible tools (explicit allowlist shown to the operator on pin), decoupled from the minting run. Ungranted pins stay read-only — still useful for display dashboards. v1 pins to the originating session's board; cross-session pinning needs a lease broker and waits. Coordinate with open PR #109807 (ui/message composer routing, theme/size propagation).

WorkBoard integration

The WorkBoard integration program keeps cards and boards plugin-owned while stitching dispatched cards back to their session boards through the existing sessionKey and runId, exposing WorkBoard feeds and dispatch through plugin-declared bindings and actions, and composing those results with the existing html and mcp-app widget kinds instead of introducing a WorkBoard-specific widget type.

Layout: fluid grid

12 columns, fixed row height, auto-compacting (gravity-up, push-aside on drag — gridstack semantics, implemented natively; grid math stays pure and DOM-free). Widget layout state per tab: { name, w (1-12), h (rows) } plus order. Agent vocabulary:

주요 항목:

size: sm (3×3) · md (6×4) · lg (8×6) · xl (12×8) · full
after: <widgetName> optional ordering anchor; omitted = append
User drags/resizes freely; the same order+size model round-trips.

Data model (per-agent DB)

New tables in agents//agent/openclaw-agent.sqlite (requires an agent-DB schema-version bump — operator sign-off required before this lands):

CREATE TABLE board_tabs (
  session_key TEXT NOT NULL,
  tab_id      TEXT NOT NULL,           -- slug
  title       TEXT NOT NULL,
  position    INTEGER NOT NULL,
  chat_dock   TEXT NOT NULL DEFAULT 'right',  -- left|right|bottom|hidden
  created_by  TEXT NOT NULL,           -- 'user' | 'agent'
  PRIMARY KEY (session_key, tab_id)
) STRICT;

CREATE TABLE board_widgets (
  session_key  TEXT NOT NULL,
  name         TEXT NOT NULL,          -- stable widget name
  tab_id       TEXT NOT NULL,
  title        TEXT,
  html         BLOB NOT NULL,          -- wrapped document source
  sha256       TEXT NOT NULL,
  revision     INTEGER NOT NULL,
  size_w       INTEGER NOT NULL,
  size_h       INTEGER NOT NULL,
  position     INTEGER NOT NULL,       -- order within tab (auto-compact input)
  manifest     TEXT NOT NULL DEFAULT '{}',  -- capability manifest JSON
  grant_state  TEXT NOT NULL DEFAULT 'none', -- none|pending|granted|rejected
  granted_sha  TEXT,                   -- byte-frozen grant
  created_by   TEXT NOT NULL,
  created_at   INTEGER NOT NULL,
  updated_at   INTEGER NOT NULL,
  PRIMARY KEY (session_key, name)
) STRICT;

Protocol surface

RPCs (core method table, typebox schemas in gateway-protocol):

주요 항목:

board.get { sessionKey } → tabs + widget metadata (no bytes) — operator.read
board.update { sessionKey, ops[] } — tab CRUD/reorder, widget move/resize/
board.widget.put { sessionKey, name, html, manifest, placement } —
board.widget.grant { sessionKey, name, decision } — operator.approvals
board.event { ticket, payload } — ticket-bound tier-1 state event ingest;
board.prompt.authorize { ticket } — returns whether a visible prompt send
board.data.read { ticket, bindingId, params? } — gateway-side allowlisted
board.action { ticket, action, ... } — exact-grant automation dispatch
board.changed { sessionKey, revision, widget? } — persisted state changed;
board.command { sessionKey, command } — transient UI drive (agent switches

Agent tools

Three tools total (core, always registered; rendering gated on the inline-widgets client cap as today):

주요 항목:

`show_widget { title, widget_code, name?, pin?, size?, tab?, after?,
dashboard { action, ... } — board management verbs: read, tab_create,
Existing cron tools cover the automation tier; no new tool needed.

What this replaces

false, never in a stable release (first appeared in 2026.7.2 betas). No migration; a doctor rule removes stale /workspaces/` if present. Harvested ideas: pure grid math, bridge security model (port bootstrap, binding gating, rate limits), byte-frozen approval.

주요 항목:

extensions/workspaces is deleted. Experimental, `enabledByDefault:
Widget hosting moves from extensions/canvas to core. The canvas doc

Non-goals (this program)

Control UI; the inline-widget path is unchanged).

주요 항목:

Multi-user board sharing/ACLs (future; will arrive via session sharing).
Native macOS/iOS board rendering (they get it wherever they embed the
Builtin data widgets (sessions/usage/cron cards) — the capability bridge plus

Implementation plan

Independent worktrees, Codex-built, review+land sequentially. Land-then-fix.

#	Branch	Scope	Depends on
T1	`claude/dashboard-remove-workspaces`	Delete workspaces plugin + UI + docs + i18n keys; doctor cleanup rule	—
T2	`claude/dashboard-canvas-core`	Promote widget hosting + `show_widget` to core; canvas plugin keeps node tool; zero behavior change	—
T3	`claude/dashboard-domain`	Agent-DB tables (schema bump), `board.*` RPCs + events, `dashboard` tool, `show_widget` pin/name/manifest args, tier-1 notices, reset-keeps-board	T2
T4	`claude/dashboard-ui`	Board face + tab strip + fluid auto-compact grid + chat dock (left/right/bottom/hidden) + transcript pin affordance + sidebar board face + reset confirm	T3 (mock-first via dev fixtures)
T5	`claude/dashboard-capabilities`	Grant store/UI + byte freezing; move `html` widgets onto the shared sandbox host; host tools (`openclaw.prompt.send/state.emit/data.read/cron.trigger`); `net` CSP; authoring shim	T3, T4
T7	`claude/dashboard-mcp-apps`	`mcp-app` content kind: pin affordance on inline app views, descriptor storage, lease re-mint/refresh, durable server-tool grants (reuses shipped MCP Apps host)	T3, T4
T6	polish	Live E2E on a scratch gateway (real keys), screenshots, fixes, user-focused `/web/dashboard` rewrite, enable-by-default review	all

실습 체크리스트

공식 문서와 로컬 버전을 대조합니다: https://docs.openclaw.ai/web/dashboard-architecture
관련 CLI는 openclaw --help 및 하위 명령 --help로 옵션을 확인합니다.
설정 변경 시 openclaw config / openclaw doctor로 유효성을 검사합니다.
Gateway·채널·플러그인 변경 후에는 필요 시 Gateway를 재시작합니다.

자주 쓰는 명령·설정 예시

CREATE TABLE board_tabs (
  session_key TEXT NOT NULL,
  tab_id      TEXT NOT NULL,           -- slug
  title       TEXT NOT NULL,
  position    INTEGER NOT NULL,
  chat_dock   TEXT NOT NULL DEFAULT 'right',  -- left|right|bottom|hidden
  created_by  TEXT NOT NULL,           -- 'user' | 'agent'
  PRIMARY KEY (session_key, tab_id)
) STRICT;

CREATE TABLE board_widgets (
  session_key  TEXT NOT NULL,
  name         TEXT NOT NULL,          -- stable widget name
  tab_id       TEXT NOT NULL,
  title        TEXT,
  html         BLOB NOT NULL,          -- wrapped document source
  sha256       TEXT NOT NULL,
  revision     INTEGER NOT NULL,
  size_w       INTEGER NOT NULL,
  size_h       INTEGER NOT NULL,
  position     INTEGER NOT NULL,       -- order within tab (auto-compact input)
  manifest     TEXT NOT NULL DEFAULT '{}',  -- capability manifest JSON
  grant_state  TEXT NOT NULL DEFAULT 'none', -- none|pending|granted|rejected
  granted_sha  TEXT,                   -- byte-frozen grant
  created_by   TEXT NOT NULL,
  created_at   INTEGER NOT NULL,
  updated_at   INTEGER NOT NULL,
  PRIMARY KEY (session_key, name)
) STRICT;

Dashboard Architecture

핵심 요약

Session dashboards: architecture and implementation plan (technical design, pre-GA)

한국어 가이드 범위: web/dashboard-architecture 경로의 설정·명령·제약·예시를 학습용으로 재구성합니다.

문서 구성

공식 문서의 주요 섹션은 다음과 같습니다.

Vision
개념
UX flows
Interaction tiers
Widget model and hosting
Widgets host content; MCP apps are one content kind
Plugin capability declarations
Modeled residual: WebRTC data channels
Transcript display: one widget card
Server-sourced widgets (pinned MCP apps)
WorkBoard integration
Layout: fluid grid
Data model (per-agent DB)
Protocol surface
Agent tools
What this replaces
Non-goals (this program)
Implementation plan

상세 내용

본문

Vision

주요 항목:

A board is a face of a session, not a new object. Every session (thread)
Agent parity. Everything the user can do on a board, the agent can do
Native, not embedded. The board is Lit components in the Control UI shell
Small agent surface. Widgets are addressed by stable name and updated in
Capabilities over trust. Widget code is arbitrary agent-authored HTML/JS

개념

Concept	Definition
Session (thread)	Existing gateway session, keyed by stable `sessionKey`. Owned by an agent.
Board	The widget face of one session. Exists iff the session has widgets/tabs. Survives `/new`/`/reset` (attached to `sessionKey`, not the transcript).
Tab	A presentation page of a board: which widgets, their arrangement, and the chat dock state (`left`/`right`/`bottom`/`hidden`). Boards start with one implicit tab.
Widget	Named, sandboxed HTML/JS program owned by the session. Addressed as `sessionKey` + `name`. Updated in place by name.
Capability manifest	Per-widget declaration of reach: `data` (read bindings), `actions` (allowlisted verbs), `prompt` (send to session), `net` (allowed origins).
Pin (widget)	Moving a transcript widget onto the session's board (user affordance or agent tool arg). Unpin removes it from the board.
Pin (session)	Existing sidebar pinning of sessions. A pinned session with a board opens on its board face.

UX flows

in the transcript exactly as today → hover shows Pin to dashboard → widget appears on the session's board. The agent can pass pin: true to do the same.

주요 항목:

Graduation: agent calls show_widget in any chat → widget renders inline
Board view: a session with a board gets a face toggle (Chat / Dashboard).
Drag: user drags widgets; grid auto-compacts (widgets float up, neighbors
Reset warning: /new / /reset on a board-bearing session asks for
Sidebar: pinned sessions render their board face when they have one.
Interactions (three tiers, see below): silent state events, visible

Interaction tiers

State events (default). Widget UI interactions the model should know about but not respond to. bridge.emitState({...}) appends a structured session notice (same mechanism as group-activity notices). No agent turn is started; the model sees accumulated notices on its next run. 2. Prompts (explicit talk). bridge.sendPrompt(text) — requires user activation; sends a visible user message into the session (the docked chat shows it). Rate-limited; each send is user-confirmed unless the widget holds the prompt capability grant. 3. Automation. bridge.runAction(name, args) — fires a manifest-declared action. Initial verb set: cron.trigger (run an existing cron job now) and binding.refresh. Cron jobs already run in visible, isolated run-sessions and can use a cheaper model: that is the "small model powers the widget" path. No hidden sessions anywhere.

주요 항목:

Inline (transcript) widgets keep the current canvas-document pipeline:
Board widgets are session state: bytes live in the owning agent's SQLite
Update in place: re-emitting a widget with the same name replaces the
Byte freezing: granted capabilities bind to the sha256 of the widget

Widgets host content; MCP apps are one content kind

The widget is the OpenClaw primitive: the named, pinned, sized, session-owned board cell with a grant record. What renders inside it is a content kind:

주요 항목:

html — agent-authored via show_widget, bytes in board storage.
mcp-app — a third-party MCP app view (ui:// resource from a configured
One sandbox host. html widgets render through the same hardened
One authorization model. A widget's reach is a granted allowlist,
Host tools for html widgets (exposed over the widget bridge, checked
openclaw.prompt.send — tier 2; routed through the visible composer,
openclaw.state.emit — tier 1 session notices (coalesced, size-capped)
openclaw.data.read — parameterized read-only bindings (existing
openclaw.cron.trigger — tier 3 automation
net = CSP. Network reach uses the already-shipped per-widget CSP
Grants. A widget declaring nothing renders immediately (sandboxed,
Authoring shim. The document wrapper injects window.openclaw.prompt,

Plugin capability declarations

Modeled residual: WebRTC data channels

Server-sourced widgets (pinned MCP apps)

WorkBoard integration

Layout: fluid grid

주요 항목:

size: sm (3×3) · md (6×4) · lg (8×6) · xl (12×8) · full
after: <widgetName> optional ordering anchor; omitted = append
User drags/resizes freely; the same order+size model round-trips.

Data model (per-agent DB)

New tables in agents//agent/openclaw-agent.sqlite (requires an agent-DB schema-version bump — operator sign-off required before this lands):

CREATE TABLE board_tabs (
  session_key TEXT NOT NULL,
  tab_id      TEXT NOT NULL,           -- slug
  title       TEXT NOT NULL,
  position    INTEGER NOT NULL,
  chat_dock   TEXT NOT NULL DEFAULT 'right',  -- left|right|bottom|hidden
  created_by  TEXT NOT NULL,           -- 'user' | 'agent'
  PRIMARY KEY (session_key, tab_id)
) STRICT;

CREATE TABLE board_widgets (
  session_key  TEXT NOT NULL,
  name         TEXT NOT NULL,          -- stable widget name
  tab_id       TEXT NOT NULL,
  title        TEXT,
  html         BLOB NOT NULL,          -- wrapped document source
  sha256       TEXT NOT NULL,
  revision     INTEGER NOT NULL,
  size_w       INTEGER NOT NULL,
  size_h       INTEGER NOT NULL,
  position     INTEGER NOT NULL,       -- order within tab (auto-compact input)
  manifest     TEXT NOT NULL DEFAULT '{}',  -- capability manifest JSON
  grant_state  TEXT NOT NULL DEFAULT 'none', -- none|pending|granted|rejected
  granted_sha  TEXT,                   -- byte-frozen grant
  created_by   TEXT NOT NULL,
  created_at   INTEGER NOT NULL,
  updated_at   INTEGER NOT NULL,
  PRIMARY KEY (session_key, name)
) STRICT;

Protocol surface

RPCs (core method table, typebox schemas in gateway-protocol):

주요 항목:

board.get { sessionKey } → tabs + widget metadata (no bytes) — operator.read
board.update { sessionKey, ops[] } — tab CRUD/reorder, widget move/resize/
board.widget.put { sessionKey, name, html, manifest, placement } —
board.widget.grant { sessionKey, name, decision } — operator.approvals
board.event { ticket, payload } — ticket-bound tier-1 state event ingest;
board.prompt.authorize { ticket } — returns whether a visible prompt send
board.data.read { ticket, bindingId, params? } — gateway-side allowlisted
board.action { ticket, action, ... } — exact-grant automation dispatch
board.changed { sessionKey, revision, widget? } — persisted state changed;
board.command { sessionKey, command } — transient UI drive (agent switches

Agent tools

Three tools total (core, always registered; rendering gated on the inline-widgets client cap as today):

주요 항목:

`show_widget { title, widget_code, name?, pin?, size?, tab?, after?,
dashboard { action, ... } — board management verbs: read, tab_create,
Existing cron tools cover the automation tier; no new tool needed.

What this replaces

주요 항목:

extensions/workspaces is deleted. Experimental, `enabledByDefault:
Widget hosting moves from extensions/canvas to core. The canvas doc

Non-goals (this program)

Control UI; the inline-widget path is unchanged).

주요 항목:

Multi-user board sharing/ACLs (future; will arrive via session sharing).
Native macOS/iOS board rendering (they get it wherever they embed the
Builtin data widgets (sessions/usage/cron cards) — the capability bridge plus

Implementation plan

Independent worktrees, Codex-built, review+land sequentially. Land-then-fix.

#	Branch	Scope	Depends on
T1	`claude/dashboard-remove-workspaces`	Delete workspaces plugin + UI + docs + i18n keys; doctor cleanup rule	—
T2	`claude/dashboard-canvas-core`	Promote widget hosting + `show_widget` to core; canvas plugin keeps node tool; zero behavior change	—
T3	`claude/dashboard-domain`	Agent-DB tables (schema bump), `board.*` RPCs + events, `dashboard` tool, `show_widget` pin/name/manifest args, tier-1 notices, reset-keeps-board	T2
T4	`claude/dashboard-ui`	Board face + tab strip + fluid auto-compact grid + chat dock (left/right/bottom/hidden) + transcript pin affordance + sidebar board face + reset confirm	T3 (mock-first via dev fixtures)
T5	`claude/dashboard-capabilities`	Grant store/UI + byte freezing; move `html` widgets onto the shared sandbox host; host tools (`openclaw.prompt.send/state.emit/data.read/cron.trigger`); `net` CSP; authoring shim	T3, T4
T7	`claude/dashboard-mcp-apps`	`mcp-app` content kind: pin affordance on inline app views, descriptor storage, lease re-mint/refresh, durable server-tool grants (reuses shipped MCP Apps host)	T3, T4
T6	polish	Live E2E on a scratch gateway (real keys), screenshots, fixes, user-focused `/web/dashboard` rewrite, enable-by-default review	all

실습 체크리스트

공식 문서와 로컬 버전을 대조합니다: https://docs.openclaw.ai/web/dashboard-architecture
관련 CLI는 openclaw --help 및 하위 명령 --help로 옵션을 확인합니다.
설정 변경 시 openclaw config / openclaw doctor로 유효성을 검사합니다.
Gateway·채널·플러그인 변경 후에는 필요 시 Gateway를 재시작합니다.

자주 쓰는 명령·설정 예시

CREATE TABLE board_tabs (
  session_key TEXT NOT NULL,
  tab_id      TEXT NOT NULL,           -- slug
  title       TEXT NOT NULL,
  position    INTEGER NOT NULL,
  chat_dock   TEXT NOT NULL DEFAULT 'right',  -- left|right|bottom|hidden
  created_by  TEXT NOT NULL,           -- 'user' | 'agent'
  PRIMARY KEY (session_key, tab_id)
) STRICT;

CREATE TABLE board_widgets (
  session_key  TEXT NOT NULL,
  name         TEXT NOT NULL,          -- stable widget name
  tab_id       TEXT NOT NULL,
  title        TEXT,
  html         BLOB NOT NULL,          -- wrapped document source
  sha256       TEXT NOT NULL,
  revision     INTEGER NOT NULL,
  size_w       INTEGER NOT NULL,
  size_h       INTEGER NOT NULL,
  position     INTEGER NOT NULL,       -- order within tab (auto-compact input)
  manifest     TEXT NOT NULL DEFAULT '{}',  -- capability manifest JSON
  grant_state  TEXT NOT NULL DEFAULT 'none', -- none|pending|granted|rejected
  granted_sha  TEXT,                   -- byte-frozen grant
  created_by   TEXT NOT NULL,
  created_at   INTEGER NOT NULL,
  updated_at   INTEGER NOT NULL,
  PRIMARY KEY (session_key, name)
) STRICT;

Dashboard Architecture

핵심 요약

문서 구성

상세 내용

본문

Vision

개념

UX flows

Interaction tiers

Widget model and hosting

Widgets host content; MCP apps are one content kind

Plugin capability declarations

Modeled residual: WebRTC data channels

Transcript display: one widget card

Server-sourced widgets (pinned MCP apps)

WorkBoard integration

Layout: fluid grid

Data model (per-agent DB)

Protocol surface

Agent tools

What this replaces

Non-goals (this program)

Implementation plan

실습 체크리스트

자주 쓰는 명령·설정 예시

관련 링크

관련 가이드

Dashboard Architecture

핵심 요약

문서 구성

상세 내용

본문

Vision

개념

UX flows

Interaction tiers

Widget model and hosting

Widgets host content; MCP apps are one content kind

Plugin capability declarations

Modeled residual: WebRTC data channels

Transcript display: one widget card

Server-sourced widgets (pinned MCP apps)

WorkBoard integration

Layout: fluid grid

Data model (per-agent DB)

Protocol surface

Agent tools

What this replaces

Non-goals (this program)

Implementation plan

실습 체크리스트

자주 쓰는 명령·설정 예시

관련 링크

관련 가이드