v0.8.11: perf improvements, instant chat navigation, code hygiene

.claude/rules/emcn-components.md

@@ -33,6 +33,16 @@ The menu surface intentionally diverges from the pill: `dropdown-menu.tsx` items
 - **`ChipTimePicker`** — minute-granular time sibling of `ChipDatePicker`, a `ChipInput` that leniently parses typed input (`9:47`, `947`, `2:05pm`, `14:30`), commits on Enter/blur, and re-renders the canonical `9:47 AM` label.
 - **`DropdownMenu`** — the canonical context/action menu (Radix-backed). Not a chip, but the standard menu for command/action lists; reach for it instead of a hand-rolled popover. Its surface intentionally diverges from the chip pill (`text-small`, `gap-2`) — keep them distinct. For a pill that opens a value picker, use `ChipDropdown`/`ChipSelect` instead.
 
+## Modal keyboard defaults
+
+Declare keyboard intent on the action-owning primitive; never add document-level or per-callsite Enter listeners.
+
+- `ChipModalFooter` defaults to `defaultAction='primary'`. A plain Enter in a canonical single-line field or a custom plain input invokes the enabled primary action. Use `'none'` when submission must require an explicit click, such as an irreversible destructive action or an editor whose nested control owns Enter. Use `'dismiss'` only when dismissal is genuinely the modal's default decision.
+- `ChipConfirmModal` fails safe with `defaultAction='dismiss'`. Opt into `'confirm'` only for an audited, low-impact reversible or non-destructive decision. Deleting an aggregate resource such as a workflow, table, knowledge base, or folder remains `'dismiss'` even when it can be restored, because the action takes a broad dependent graph offline. Use `'none'` for typed confirmations and severe account, ownership, or access changes. Button color never determines keyboard behavior.
+- Textareas, native forms, buttons, links, comboboxes, menus, listboxes, tag/email inputs, IME composition, modified Enter, and disabled or pending actions retain their native behavior. A native form remains the sole submission path so browser validation is not bypassed.
+- A custom field containing a search, token editor, or another input that owns Enter must set `submitOnEnter={false}` on `ChipModalField`. Do not attach a duplicate `onKeyDown` handler merely to call the footer action.
+- Initial focus goes to the first visible editable text control. With no text control, the declared real button receives focus; `'none'` focuses the dialog surface. A safe dismiss default never turns Enter in a text field into data loss—the field simply does not publish a submit action.
+
 ## Authoring principles
 
 - **One source of truth for shared chrome.** Compose from `chip-chrome.ts` / `chipVariants`; never duplicate the chrome string.

.claude/rules/sim-react-performance.md

@@ -90,6 +90,19 @@ const [{ id }, { kbName }] = await Promise.all([params, searchParams])
 
 Only keep awaits sequential when a later call genuinely uses an earlier result, or when the ordering is deliberate (rate-limited batches, retry loops, write-then-read).
 
+## Prefetch dynamic destination lists on intent
+
+For long lists of dynamic destinations, do not viewport-prefetch every row and do not assume
+`router.prefetch()` warms the full route: in Next 16 it uses the automatic/PPR strategy. Gate
+`<Link prefetch={true}>` behind deliberate hover or keyboard focus, and prefetch destination
+server state with the consumer's shared React Query options. A short, cancelable hover dwell
+avoids drive-by downloads. Do not treat `touchstart` as intent because it also begins scrolling;
+let the actual unmodified click start the data request.
+
+If a continuity-focused surface intentionally omits `loading.tsx` so the current view remains
+mounted until its peer is ready, the intent path must warm both the full route and its critical
+data. Otherwise keep the loading boundary so dynamic navigation remains responsive.
+
 ## Local feature barrels are the convention — do not "fix" them
 
 Tooling (e.g. react-doctor's `no-barrel-import`) will flag imports from local `index.ts` barrels as a bundle cost. In this repo that is a **false positive**: barrel imports for 3+ export folders are mandated by `.claude/rules/sim-imports.md`. Leave them.

.claude/rules/sim-settings-pages.md

@@ -13,8 +13,9 @@ The Next.js `settings/[section]/layout.tsx` owns all settings page chrome via
 `SettingsHeaderShell` — a fixed header bar (a left back chip + right-aligned
 action chips), a scroll region, and a centered `max-w-[48rem]` content column led
 by a **title + description from navigation metadata**. The chrome stays mounted
-across section navigation (it never re-renders or re-lays-out). Each section
-renders through the **`SettingsPanel`** registrar
+across section navigation. Its routed title and description are available before
+the section body resolves. Each section renders through the **`SettingsPanel`**
+registrar
 (`@/app/workspace/[workspaceId]/settings/components/settings-panel`), which feeds
 the shell its header data and renders only the section body. Sections supply
 **data**, never chrome.
@@ -82,6 +83,9 @@ return (
   `children` instead and omit the prop.
 - `title?` / `description?` — overrides for the nav-driven defaults. **Only** for a
   detail sub-view that needs a different heading; normal pages never pass these.
+  A top-level page's header identity must remain stable while its data loads:
+  never replace navigation metadata with client-fetched copy after first paint.
+  Put data-dependent context in the page body instead.
 - `scrollContainerRef?: React.Ref<HTMLDivElement>` — forwards a ref to the scroll
   region (e.g. programmatic scroll-to-bottom).
 

.claude/rules/sim-url-state.md

@@ -143,6 +143,11 @@ import KnowledgeBaseLoading from '@/app/workspace/[workspaceId]/knowledge/[id]/l
 
 Reference: `apps/sim/app/workspace/[workspaceId]/knowledge/[id]/page.tsx`.
 
+The narrow exception is a continuity-focused peer switch that deliberately keeps the current
+view mounted and follows the full-route plus critical-data intent-prefetch rule in
+`sim-react-performance.md`. It still needs a real in-page Suspense fallback; it only omits the
+route-level `loading.tsx` that would replace the current peer before the destination is ready.
+
 This applies to **page entries**. An inner `<Suspense>` wrapping a `lazy()` component is the exception: there `fallback={null}` is correct, precisely so the suspend resolves at the nearest boundary instead of flashing the whole route — see `sim-imports.md`, "Code-splitting through barrels".
 
 ## Debounced text inputs

.github/workflows/test-build.yml

@@ -14,8 +14,18 @@ jobs:
     timeout-minutes: 15
 
     steps:
+      # The diff-based audits below need a base commit to read, and the default
+      # depth of 1 clones a single commit with no parent. They normally fetch
+      # their base by SHA (see "Resolve base ref"), so this depth only covers the
+      # `HEAD~1` fallback — but without it that fallback resolves to nothing.
+      #
+      # Worth stating because the failure was invisible for so long: the migration
+      # audit read the resulting `git diff` failure as "no migrations changed" and
+      # exited 0, so it had never actually run on a push build.
       - name: Checkout code
         uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10 # v6
+        with:
+          fetch-depth: 2
 
       - name: Setup Bun
         uses: oven-sh/setup-bun@0c5077e51419868618aeaa5fe8019c62421857d6 # v2
@@ -104,15 +114,40 @@ jobs:
 
           echo "✅ All env flags are properly configured"
 
-      - name: Check block registry invariants
+      # One fetch for both base-ref audits, and no `|| true`: a swallowed fetch leaves
+      # the base ref absent, which neither audit can tell apart from a branch that
+      # changed nothing. The block-registry check at least degrades to a visible
+      # `⚠ … skipping` line; the migration audit printed `✓ No new migrations to
+      # check` and exited 0, clearing the only guard on production DDL.
+      #
+      # Depth stays at 1 — without a merge-base the migration audit diffs the two
+      # tips, which under `--diff-filter=AM` is exactly the migrations new here.
+      # Resolved once for both diff-based audits, and never with `|| true`: a
+      # swallowed fetch leaves the base absent, which neither audit can tell apart
+      # from a branch that changed nothing.
+      #
+      # On push the base is `github.event.before`, the tip the branch had before
+      # this push — not `HEAD~1`, which names only the last commit and would let a
+      # multi-commit push slip every earlier commit's migrations past the audit.
+      # It is fetched by SHA at depth 1; the audits diff two tips and need no
+      # common ancestry. An all-zero `before` means the branch is new and has no
+      # predecessor to diff, so `HEAD~1` remains the fallback there.
+      - name: Resolve base ref for diff-based audits
+        id: audit_base
         run: |
           if [ "${{ github.event_name }}" = "pull_request" ]; then
-            BASE_REF="origin/${{ github.base_ref }}"
-            git fetch --depth=1 origin "${{ github.base_ref }}" 2>/dev/null || true
+            git fetch --depth=1 origin "${{ github.base_ref }}"
+            echo "ref=origin/${{ github.base_ref }}" >> "$GITHUB_OUTPUT"
+          elif [ -n "${{ github.event.before }}" ] &&
+               [ "${{ github.event.before }}" != "0000000000000000000000000000000000000000" ]; then
+            git fetch --depth=1 origin "${{ github.event.before }}"
+            echo "ref=${{ github.event.before }}" >> "$GITHUB_OUTPUT"
           else
-            BASE_REF="HEAD~1"
+            echo "ref=HEAD~1" >> "$GITHUB_OUTPUT"
           fi
-          bun run apps/sim/scripts/check-block-registry.ts "$BASE_REF"
+
+      - name: Check block registry invariants
+        run: bun run apps/sim/scripts/check-block-registry.ts "${{ steps.audit_base.outputs.ref }}"
 
       - name: Lint code
         run: bun run lint:check
@@ -127,14 +162,7 @@ jobs:
         run: bun run docs-manifest:check
 
       - name: Migration safety (zero-downtime) audit
-        run: |
-          if [ "${{ github.event_name }}" = "pull_request" ]; then
-            BASE_REF="origin/${{ github.base_ref }}"
-            git fetch --depth=1 origin "${{ github.base_ref }}" 2>/dev/null || true
-          else
-            BASE_REF="HEAD~1"
-          fi
-          bun run check:migrations "$BASE_REF"
+        run: bun run check:migrations "${{ steps.audit_base.outputs.ref }}"
 
       # Every workspace, not just realtime. packages/emcn, packages/utils,
       # apps/desktop and apps/docs had no type check in CI at all; apps/sim's

README.md

@@ -88,6 +88,7 @@ npx sim-setup add sandbox
 npx sim-setup add jobs
 npx sim-setup add cache
 npx sim-setup add knowledge
+npx sim-setup add chat
 npx sim-setup add llm
 npx sim-setup add integration slack
 ```

apps/docs/components/workflow-preview/docs-container-node.tsx

@@ -8,6 +8,7 @@ interface DocsContainerData {
   name: string
   blockType: string
   size?: { width: number; height: number }
+  parentId?: string
 }
 
 /**
@@ -24,6 +25,7 @@ export const DocsContainerNode = memo(function DocsContainerNode({
     name: data.name,
     width: data.size?.width,
     height: data.size?.height,
+    parentId: data.parentId,
     isPreview: true,
   }
 

apps/docs/components/workflow-preview/workflow-data.test.ts

@@ -0,0 +1,63 @@
+/**
+ * @vitest-environment node
+ */
+import { BLOCK_Z_BASE, CONTAINER_CHILD_Z_BASE, getEdgeZIndex } from '@sim/workflow-renderer'
+import { describe, expect, it } from 'vitest'
+import { type PreviewBlock, type PreviewWorkflow, toReactFlowElements } from './workflow-data'
+
+const block = (
+  overrides: Partial<PreviewBlock> & Pick<PreviewBlock, 'id' | 'type'>
+): PreviewBlock => ({
+  name: overrides.id,
+  bgColor: '#000000',
+  rows: [],
+  position: { x: 0, y: 0 },
+  ...overrides,
+})
+
+const workflow: PreviewWorkflow = {
+  id: 'nested-subflows',
+  name: 'Nested subflows',
+  blocks: [
+    block({ id: 'start', type: 'starter' }),
+    block({ id: 'loop', type: 'loop', size: { width: 500, height: 300 } }),
+    block({
+      id: 'parallel',
+      type: 'parallel',
+      parentId: 'loop',
+      position: { x: 24, y: 64 },
+      size: { width: 400, height: 200 },
+    }),
+    block({ id: 'agent', type: 'agent', parentId: 'loop', position: { x: 24, y: 140 } }),
+  ],
+  edges: [
+    { id: 'start-loop', source: 'start', target: 'loop' },
+    { id: 'loop-parallel', source: 'loop', target: 'parallel' },
+    { id: 'loop-agent', source: 'loop', target: 'agent' },
+  ],
+}
+
+describe('toReactFlowElements layering', () => {
+  it('places incoming edges on their container target layer', () => {
+    const { nodes, edges } = toReactFlowElements(workflow, false, {
+      highlightEdge: 'loop-parallel',
+    })
+    const nodeById = new Map(nodes.map((node) => [node.id, node]))
+    const edgeById = new Map(edges.map((edge) => [edge.id, edge]))
+
+    expect(nodeById.get('loop')?.zIndex).toBe(0)
+    expect(nodeById.get('parallel')?.zIndex).toBe(1)
+    expect(edgeById.get('start-loop')?.zIndex).toBe(0)
+    expect(edgeById.get('loop-parallel')?.zIndex).toBe(1)
+  })
+
+  it('keeps ordinary cards above normally layered edges', () => {
+    const { nodes, edges } = toReactFlowElements(workflow)
+    const nodeById = new Map(nodes.map((node) => [node.id, node]))
+    const edgeById = new Map(edges.map((edge) => [edge.id, edge]))
+
+    expect(nodeById.get('start')?.zIndex).toBe(BLOCK_Z_BASE)
+    expect(nodeById.get('agent')?.zIndex).toBe(CONTAINER_CHILD_Z_BASE)
+    expect(edgeById.get('loop-agent')?.zIndex).toBe(getEdgeZIndex(0))
+  })
+})

apps/docs/components/workflow-preview/workflow-data.ts

@@ -1,3 +1,9 @@
+import {
+  BLOCK_Z_BASE,
+  CONTAINER_CHILD_Z_BASE,
+  getEdgeZIndex,
+  getEdgeZIndexForTarget,
+} from '@sim/workflow-renderer'
 import { type Edge, type Node, Position } from 'reactflow'
 
 /**
@@ -61,6 +67,24 @@ export interface HighlightOptions {
   selectedBlock?: string
 }
 
+/** Semantic container depth used for z-order while docs positions stay flattened. */
+function getNestingDepth(block: PreviewBlock, blocksById: Map<string, PreviewBlock>): number {
+  let depth = 0
+  let parentId = block.parentId
+  const visited = new Set<string>()
+
+  while (parentId && !visited.has(parentId)) {
+    const parent = blocksById.get(parentId)
+    if (!parent) break
+
+    visited.add(parentId)
+    depth += 1
+    parentId = parent.parentId
+  }
+
+  return depth
+}
+
 /**
  * Converts a {@link PreviewWorkflow} to React Flow nodes and edges.
  *
@@ -81,6 +105,7 @@ export function toReactFlowElements(
 
   const nodes: Node[] = workflow.blocks.map((block, index) => {
     const isContainer = Boolean(block.size)
+    const nestingDepth = getNestingDepth(block, blocksById)
     // Nested blocks are authored relative to their container; render them at
     // absolute coordinates (not React Flow parentNode children) so the edges
     // between a container and its nested blocks render reliably and on top.
@@ -92,7 +117,7 @@ export function toReactFlowElements(
       id: block.id,
       type: isContainer ? 'previewContainer' : 'previewBlock',
       position,
-      zIndex: isContainer ? 0 : 1,
+      zIndex: isContainer ? nestingDepth : block.parentId ? CONTAINER_CHILD_Z_BASE : BLOCK_Z_BASE,
       ...(block.size ? { style: { width: block.size.width, height: block.size.height } } : {}),
       data: {
         name: block.name,
@@ -103,6 +128,7 @@ export function toReactFlowElements(
         tools: block.tools,
         hideTargetHandle: block.hideTargetHandle,
         size: block.size,
+        parentId: block.parentId,
         index,
         animate,
         isHighlighted: highlightBlock === block.id || selectedBlock === block.id,
@@ -127,6 +153,14 @@ export function toReactFlowElements(
     // so edges into and out of Loop/Parallel containers still connect.
     const sourceBlock = blocksById.get(e.source)
     const targetBlock = blocksById.get(e.target)
+    const parentContainer = blocksById.get(sourceBlock?.parentId ?? targetBlock?.parentId ?? '')
+    const baseZIndex = getEdgeZIndex(
+      parentContainer ? getNestingDepth(parentContainer, blocksById) : undefined,
+      { isHighlighted: isEdgeHighlight }
+    )
+    const targetContainerZIndex = targetBlock?.size
+      ? getNestingDepth(targetBlock, blocksById)
+      : undefined
     const sourceHandle =
       e.sourceHandle ?? (sourceBlock?.size ? `${sourceBlock.type}-end-source` : 'source')
     const targetHandle = targetBlock?.size ? undefined : 'target'
@@ -142,6 +176,7 @@ export function toReactFlowElements(
       },
       sourceHandle,
       targetHandle,
+      zIndex: getEdgeZIndexForTarget(baseZIndex, targetContainerZIndex),
       data: {
         animate,
         delay: animate ? sourceIndex * BLOCK_STAGGER + BLOCK_STAGGER : 0,

apps/docs/content/docs/en/platform/credentials.mdx

@@ -119,12 +119,27 @@ Click **Details** on any secret row to open its detail view.
 From here you can:
 
 - View the **Key** and edit the **Value**
+- Toggle **Visibility** — show the value unmasked in run output; see [Visibility](#visibility)
 - Edit the **Description** — an optional note telling teammates what the secret is for. Workspace secrets only; a personal secret is not shared, so it has none
 - Manage **Members** — invite teammates by email and assign them an **Admin** or **Member** role
 - Open **See usage** — where this secret has actually been used
 
 Click **Save** to apply changes, or **Back** to return to the list.
 
+### Visibility
+
+By default, a secret's resolved value is masked everywhere Sim shows run output (see [Execution log protection](#execution-log-protection)). For values that aren't actually sensitive — a staging key, a shared base URL — that masking makes your own logs harder to read.
+
+**Show value in logs and Chat** turns masking off for one workspace secret. With it on:
+
+- Run logs, Chat, and code output show the real value instead of `{{KEY}}`
+- Files a run writes with the value in them stay readable and attachable
+- The Secrets API list includes the value for this secret, so external agents can read it directly instead of scraping logs
+
+The value becomes visible to **anyone who can see this workspace's runs** — including publicly shared log links and log exports, and regardless of member restrictions on the secret itself. Only turn it on for values you'd be comfortable printing in a log.
+
+The switch applies to future runs only. Logs written while the secret was masked stay masked, and anything written while it was visible keeps the value even if you turn masking back on. If another secret holds the same value, that value stays masked — masking always wins a conflict. Workspace secrets only; the same people who can edit the description can flip it.
+
 ### See usage
 
 **See usage** lists the runs that resolved this secret: when it was last used, what used it (a workflow, the Sim agent, or an MCP server), how it was triggered, who it resolved under, and a link to the most recent run in Logs. Rows are grouped by day, so a workflow on a schedule reads as one row per day rather than thousands.

apps/docs/content/docs/en/platform/self-hosting/docker.mdx

@@ -134,6 +134,26 @@ OLLAMA_URL=http://192.168.1.100:11434 docker compose -f docker-compose.prod.yml
   Inside Docker, `localhost` refers to the container, not your host. Use `host.docker.internal` or your host's IP.
 </Callout>
 
+### LM Studio
+
+[LM Studio exposes an OpenAI-compatible API](https://lmstudio.ai/docs/developer/openai-compat). Start its local server, load a model, and enable **Serve on Local Network** so the Docker container can reach it. [Enable API authentication](https://lmstudio.ai/docs/developer/core/authentication), then set the endpoint and token in the `.env` file next to your Compose file:
+
+```bash
+# macOS/Windows
+VLLM_BASE_URL=http://host.docker.internal:1234
+
+# Linux - use your host IP instead
+# VLLM_BASE_URL=http://192.168.1.100:1234
+
+VLLM_API_KEY=your_lm_studio_api_token
+```
+
+Both the server root shown above and a URL ending in `/v1` are accepted. After recreating the `simstudio` service, its models appear in the model picker with a `vllm/` prefix; Sim removes that prefix before sending the model identifier to LM Studio.
+
+```bash
+docker compose -f docker-compose.ollama.yml up -d --force-recreate simstudio
+```
+
 ## Commands
 
 ```bash