feat: user-facing vocabulary says step, not node (goal 0053) (#110)

* feat: user-facing vocabulary says step, not node (goal 0053) -- seed copy drops internal doc refs

Sweeps every user-visible surface (locale JSON, NodeType/ConfigField
Label/Description/Output, validation and run-failure messages, seeded
workflow descriptions) from "node" to "step" per goal 0053, leaving
code identifiers (NodeType, composition.Node, RefKind wire values)
untouched. Rides the goal 0044 dry-run gap #5 rider: seeded
descriptions in builtinworkflows*.go (plus two more found in
list/builtin.go and httprequest/builtin.go) no longer cite internal
docs/adr/goal references, and check-ui-copy.sh now gates
builtinworkflows*.go the same way it already gates locale JSON.

Bumps SeedRevision (1 -> 2) on every seeded entity whose Description
changed and updates seed_fingerprints.json to match.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FJ8wStsHyu7XPLTspNjMnQ

* docs: goal 0053 acceptance recorded; ui-copy gate covers list/http seed descriptions

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FJ8wStsHyu7XPLTspNjMnQ

---------

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>

Author: alicoding

frontend/src/composition/validationCopy.test.ts

@@ -18,31 +18,31 @@ function t(key: string, vars: Record<string, unknown> = {}): string {
 }
 
 describe('formatIssuesForCopy', () => {
-  it('names the workflow, its id, counts, and each issue with its node/edge id', () => {
+  it('names the workflow, its id, counts, and each issue with its step/edge id', () => {
     const out = formatIssuesForCopy(t, 'Load sample HTML', 'load-sample-html-workflow', [
       issue({ Severity: Severity.SeverityError, NodeID: 'n1', Message: 'a workflow must start with a Trigger step' }),
       issue({ Severity: Severity.SeverityWarning, EdgeID: 'e7', Message: 'dangling edge' }),
     ])
     expect(out).toBe(
       'Mill workflow "Load sample HTML" (id: load-sample-html-workflow) — validation issues (1 error · 1 warning):\n' +
-      '- [error] node n1: a workflow must start with a Trigger step\n' +
+      '- [error] step n1: a workflow must start with a Trigger step\n' +
       '- [warning] edge e7: dangling edge',
     )
   })
 
-  it('never double-prefixes when the message already carries its node location', () => {
-    // ValidateGraph's real messages lead with "node <id>: " -- caught
-    // from an actual paste that read "node X: node X: ..." (goal 0021).
+  it('never double-prefixes when the message already carries its step location', () => {
+    // ValidateGraph's real messages lead with "step <id>: " -- guards
+    // against reading "step X: step X: ...".
     const out = formatIssuesForCopy(t, 'W', 'w-id', [
-      issue({ NodeID: 'n1', Message: 'node n1: a workflow must start with a Trigger step' }),
+      issue({ NodeID: 'n1', Message: 'step n1: a workflow must start with a Trigger step' }),
     ])
-    expect(out).toContain('- [error]: node n1: a workflow must start with a Trigger step')
-    expect(out).not.toContain('node n1: node n1:')
+    expect(out).toContain('- [error]: step n1: a workflow must start with a Trigger step')
+    expect(out).not.toContain('step n1: step n1:')
   })
 
-  it('omits the location fragment when an issue has no node or edge id', () => {
-    const out = formatIssuesForCopy(t, 'W', 'w-id', [issue({ Message: 'graph has no nodes' })])
-    expect(out).toContain('- [error]: graph has no nodes')
+  it('omits the location fragment when an issue has no step or edge id', () => {
+    const out = formatIssuesForCopy(t, 'W', 'w-id', [issue({ Message: 'graph has no steps' })])
+    expect(out).toContain('- [error]: graph has no steps')
   })
 
   it('pluralizes counts', () => {

frontend/src/composition/validationCopy.ts

@@ -18,10 +18,10 @@ export function formatIssuesForCopy(t: (key: string, opts?: Record<string, unkno
 
   const lines = issues.map((i) => {
     // ValidateGraph's own Message usually already leads with
-    // "node <id>: "/"edge <id>: " -- adding our own location fragment
-    // then double-prefixed every line (caught from a real paste, goal
-    // 0021). Only add a location when the message doesn't carry one.
-    const alreadyLocated = i.Message.startsWith('node ') || i.Message.startsWith('edge ')
+    // "step <id>: "/"edge <id>: " -- adding our own location fragment
+    // then double-prefixed every line. Only add a location when the
+    // message doesn't carry one.
+    const alreadyLocated = i.Message.startsWith('step ') || i.Message.startsWith('edge ')
     const where = alreadyLocated ? '' : i.NodeID ? t('validationCopy.nodeLocation', { id: i.NodeID }) : i.EdgeID ? t('validationCopy.edgeLocation', { id: i.EdgeID }) : ''
     return `- [${i.Severity}]${where}: ${i.Message}`
   })

frontend/src/locales/en/app.json

@@ -55,9 +55,9 @@
     "unrecognizedTitle": "Couldn't read that as a Mill export",
     "back": "Back",
     "summaryUpdate": "This will UPDATE \"{{label}}\" — replacing the current draft (the published version stays untouched until you publish).",
-    "summaryCreate": "This will CREATE \"{{label}}\" ({{nodeCount}} node{{plural}}).",
+    "summaryCreate": "This will CREATE \"{{label}}\" ({{nodeCount}} step{{plural}}).",
     "unresolvedTitle": "{{count}} reference{{plural}} won't resolve here",
-    "unresolvedItemPrefix": "node",
+    "unresolvedItemPrefix": "step",
     "unresolvedItemSuffix": "references \"{{value}}\", which doesn't exist here — point it at one before running",
     "applyFailedTitle": "Apply failed",
     "cancel": "Cancel",

frontend/src/locales/en/composition.json

@@ -27,7 +27,7 @@
     "runButtonTooltip": "Runs the saved draft (test run).",
     "run": "Run",
     "runInStepModeAriaLabel": "Run in step mode",
-    "runInStepModeTooltip": "Pauses before every node so you can inspect and edit its data."
+    "runInStepModeTooltip": "Pauses before every step so you can inspect and edit its data."
   },
   "canvasNodeView": {
     "runStatus": {
@@ -159,18 +159,18 @@
     "plainSuffix": ".",
     "breakpointSet": "Breakpoint set",
     "noBreakpoint": "No breakpoint",
-    "breakpointHint": "{{state}} — click the dot on the node card to {{action}}.",
+    "breakpointHint": "{{state}} — click the dot on the step card to {{action}}.",
     "removeIt": "remove it",
     "addOne": "add one",
     "breakpointBadge": "Breakpoint",
     "breakpointDescription": "A run pauses here to let you inspect and edit its data before it continues -- a debugging aid, not policy."
   },
   "nodeInspector": {
-    "nodeType": "Node type",
+    "nodeType": "Step type",
     "saveBeforeHotkey": "Save this workflow before assigning a hotkey.",
     "setShortcut": "Set shortcut",
     "openAccessibilitySettings": "Open Accessibility Settings",
-    "noConfiguration": "This node type takes no configuration.",
+    "noConfiguration": "This step type takes no configuration.",
     "generateTestPayload": "Generate test payload",
     "hoverToPreviewChild": "Hover to preview the child — click Open to edit it",
     "defaultOption": "(default)"
@@ -181,11 +181,11 @@
   },
   "draftWorkflowSchema": {
     "needsLabel": "A workflow needs a label",
-    "needsAtLeastOneNode": "A workflow needs at least one node",
-    "danglingConnection": "A connection references a node that no longer exists.",
-    "terminalCannotHaveOutgoing": "A Decision node (a terminal outcome) cannot have an outgoing connection.",
-    "onlyBranchCanFanOut": "Only a Branch node can have more than one outgoing connection.",
-    "needsExactlyOneStart": "A workflow must have exactly one starting node.",
+    "needsAtLeastOneNode": "A workflow needs at least one step",
+    "danglingConnection": "A connection references a step that no longer exists.",
+    "terminalCannotHaveOutgoing": "A Decision step (a terminal outcome) cannot have an outgoing connection.",
+    "onlyBranchCanFanOut": "Only a Branch step can have more than one outgoing connection.",
+    "needsExactlyOneStart": "A workflow must have exactly one starting step.",
     "notValidYetFallback": "This workflow is not valid yet."
   },
   "validationPanel": {
@@ -197,7 +197,7 @@
   },
   "validationCopy": {
     "header": "Mill workflow \"{{label}}\" (id: {{id}}) — validation issues ({{counts}}):",
-    "nodeLocation": " node {{id}}",
+    "nodeLocation": " step {{id}}",
     "edgeLocation": " edge {{id}}"
   },
   "workflowEditorTab": {
@@ -214,7 +214,7 @@
   },
   "nodePalette": {
     "addSteps": "Add steps",
-    "onlyOneTriggerTitle": "A workflow can only have one trigger. Select the existing trigger node on the canvas to change its type instead.",
+    "onlyOneTriggerTitle": "A workflow can only have one trigger. Select the existing trigger step on the canvas to change its type instead.",
     "searchPlaceholder": "Search steps…",
     "searchAriaLabel": "Search steps",
     "noMatches": "No steps match \"{{query}}\""
@@ -245,11 +245,11 @@
     "redoAriaLabel": "Redo",
     "autoLayoutAriaLabel": "Auto-layout",
     "deleteSelectedAriaLabel": "Delete selected",
-    "addStepsHint": "Add steps to drag a node type onto the canvas, connect them, click a node to configure it."
+    "addStepsHint": "Add steps to drag a step type onto the canvas, connect them, click a step to configure it."
   },
   "compositionCanvas": {
-    "onlyDecisionEdgesConfigurable": "Only a Decision node’s outgoing edges are configurable.",
-    "selectNodeToConfigure": "Select a node to configure it."
+    "onlyDecisionEdgesConfigurable": "Only a Decision step’s outgoing edges are configurable.",
+    "selectNodeToConfigure": "Select a step to configure it."
   },
   "compositionView": {
     "heading": "Workflows",

frontend/src/locales/en/configure.json

@@ -105,7 +105,7 @@
     "deleteAriaLabel": "Delete {{label}}",
     "searchPlaceholder": "Search lists…",
     "emptyHeading": "No lists yet",
-    "emptyDescription": "A reusable typed dataset a workflow's List Search (or List Lookup) node can resolve against.",
+    "emptyDescription": "A reusable typed dataset a workflow's List Search (or List Lookup) step can resolve against.",
     "columnsRowsSummary": "{{columns}} columns, {{rows}} rows",
     "deleteConfirmTitle": "Delete list?",
     "deleteConfirmBody": "This permanently deletes \"{{label}}\". This cannot be undone.",
@@ -158,21 +158,21 @@
     "deleteAriaLabel": "Delete {{label}}",
     "searchPlaceholder": "Search execution environments…",
     "emptyHeading": "No execution environments yet",
-    "emptyDescription": "A reusable, pinned shell/directory/env a code-execution workflow node can run inside.",
+    "emptyDescription": "A reusable, pinned shell/directory/env a code-execution workflow step can run inside.",
     "freshTempDirPerRun": "fresh temp dir per run",
     "deleteConfirmTitle": "Delete execution environment?",
     "deleteConfirmBody": "This permanently deletes \"{{label}}\". This cannot be undone."
   },
   "configureDecisions": {
     "heading": "Decisions",
     "newDecision": "New decision",
-    "pageDescription": "A Decision is a reusable, typed terminal outcome that a workflow's Decision node reaches to end the run with a real category and typed result, instead of just running out of steps.",
+    "pageDescription": "A Decision is a reusable, typed terminal outcome that a workflow's Decision step reaches to end the run with a real category and typed result, instead of just running out of steps.",
     "label": "Label",
     "category": "Category",
     "categoryCaptionEditing": "Cannot be changed after creation -- duplicate this Decision to create one with a different category.",
     "categoryCaptionCreating": "Cannot be changed after creation -- duplicate this Decision later to create one with a different category.",
     "outputs": "Outputs",
-    "outputsDescription": "This Decision's typed result fields, bound by a workflow's decision-outcome node when it reaches this outcome.",
+    "outputsDescription": "This Decision's typed result fields, bound by a workflow's decision-outcome step when it reaches this outcome.",
     "keyPlaceholder": "key",
     "labelPlaceholder": "label",
     "enumValuesPlaceholder": "enum values, comma separated",
@@ -194,7 +194,7 @@
     "deleteAriaLabel": "Delete {{label}}",
     "searchPlaceholder": "Search decisions…",
     "emptyHeading": "No decisions yet",
-    "emptyDescription": "A reusable, typed TERMINAL outcome a workflow's Decision node reaches to end the run.",
+    "emptyDescription": "A reusable, typed TERMINAL outcome a workflow's Decision step reaches to end the run.",
     "outputsSummary": "Outputs: {{keys}}",
     "none": "none",
     "deleteConfirmTitle": "Delete decision?",
@@ -223,7 +223,7 @@
     "deleteAriaLabel": "Delete {{label}}",
     "searchPlaceholder": "Search MCP servers…",
     "emptyHeading": "No MCP servers yet",
-    "emptyDescription": "A reusable stdio connection an mcp-tool-call workflow node can resolve by ID.",
+    "emptyDescription": "A reusable stdio connection an mcp-tool-call workflow step can resolve by ID.",
     "serverTools": "{{label}} — tools",
     "noToolsExposed": "This server exposes no tools.",
     "deleteConfirmTitle": "Delete MCP server?",
@@ -318,7 +318,7 @@
     "requestBody": "Request body",
     "addBodyField": "Add body field",
     "outputHeading": "Output — response schema",
-    "outputDescription": "The attributes this request's response provides, for a workflow node to read and bind into its own Attributes.",
+    "outputDescription": "The attributes this request's response provides, for a workflow step to read and bind into its own Attributes.",
     "addOutputField": "Add output field",
     "fieldNameAriaLabel": "Field name",
     "namePlaceholder": "name",
@@ -434,7 +434,7 @@
     "removeHeaderAriaLabel": "Remove header",
     "addHeader": "Add header",
     "schema": "Schema",
-    "schemaDescription": "The payload's structure only: typed input/output fields a workflow node can bind Attributes to. Method and URL live above and are never part of the schema. Leave empty to keep using a literal request body.",
+    "schemaDescription": "The payload's structure only: typed input/output fields a workflow step can bind Attributes to. Method and URL live above and are never part of the schema. Leave empty to keep using a literal request body.",
     "multiOperationNote": "This request's stored schema declares {{count}} operations, but a request is one call. Remove the extras below, or duplicate the request once per operation.",
     "hideRawOpenapi": "Hide raw OpenAPI",
     "viewRawOpenapi": "View raw OpenAPI",

internal/domain/composition/aiclassify.go

@@ -53,7 +53,7 @@ func init() {
 		Effect: guardrail.ClassExternal, // dynamic downgrade for a loopback provider -- aiprovider.go's aiNodeEffectOverride
 		Output: "unchanged payload -- the chosen category is written into the named Attribute below",
 		Label:  "AI: Classify",
-		Description: "Sends the running payload (plus an optional instruction) to a Configure-authored AI provider and asks it to pick exactly one of this step's own declared categories, writing the choice into a named Attribute -- a dedicated classification node (Dify's own precedent researched in docs/goals/0031-ai-node-family.md), not composed from extract-structured, since \"pick one of these labels\" is common enough across real workflows to deserve its own shape. Pairs with Branch: route on the written Attribute to act on the classification. The category list is this workflow's own business decision (node-local, not Configure-authored) -- two workflows classifying into different category sets is normal, not drift.",
+		Description: "Sends the running payload (plus an optional instruction) to a Configure-authored AI provider and asks it to pick exactly one of this step's own declared categories, writing the choice into a named Attribute -- a dedicated classification step (Dify's own precedent researched in docs/goals/0031-ai-node-family.md), not composed from extract-structured, since \"pick one of these labels\" is common enough across real workflows to deserve its own shape. Pairs with Branch: route on the written Attribute to act on the classification. The category list is this workflow's own business decision (step-local, not Configure-authored) -- two workflows classifying into different category sets is normal, not drift.",
 		ConfigFields: []ConfigField{
 			{
 				Key: aiProviderIDConfigKey, Label: "AI provider",

internal/domain/composition/aicompletion.go

@@ -33,7 +33,7 @@ func init() {
 		Effect: guardrail.ClassExternal, // dynamic downgrade for a loopback provider -- aiprovider.go's aiNodeEffectOverride, dispatched via EffectForNode
 		Output: "the AI completion text, replacing the payload",
 		Label:  "AI: Completion",
-		Description: "Sends a prompt (plus the running payload) to a Configure-authored AI provider -- local Ollama, or a BYO OpenAI-compatible/Anthropic endpoint -- and replaces the payload with its text completion. Composition (docs/goals/0031-ai-node-family.md): the system prompt is this node's own System prompt field; the user message is Prompt followed by the current payload (blank-line separated) when the payload is non-empty, else Prompt alone. One deterministic call per run -- never a loop or autonomous agent behavior (docs/SPEC.md §1.1's locked invariant). A local (localhost/127.0.0.1/::1) provider runs without an approval ask; any other endpoint asks by default, the same posture integration-http already has for outbound calls.",
+		Description: "Sends a prompt (plus the running payload) to a Configure-authored AI provider -- local Ollama, or a BYO OpenAI-compatible/Anthropic endpoint -- and replaces the payload with its text completion. Composition (docs/goals/0031-ai-node-family.md): the system prompt is this step's own System prompt field; the user message is Prompt followed by the current payload (blank-line separated) when the payload is non-empty, else Prompt alone. One deterministic call per run -- never a loop or autonomous agent behavior (docs/SPEC.md §1.1's locked invariant). A local (localhost/127.0.0.1/::1) provider runs without an approval ask; any other endpoint asks by default, the same posture integration-http already has for outbound calls.",
 		ConfigFields: []ConfigField{
 			{
 				Key: aiProviderIDConfigKey, Label: "AI provider",

internal/domain/composition/aiextractstructured.go

@@ -121,7 +121,7 @@ func init() {
 		Effect: guardrail.ClassExternal, // dynamic downgrade for a loopback provider -- aiprovider.go's aiNodeEffectOverride
 		Output: "unchanged payload -- the extracted typed result is written into the named Attributes below",
 		Label:  "AI: Extract structured data",
-		Description: "Sends a prompt (plus the running payload) to a Configure-authored AI provider, requests a JSON-schema-constrained structured response, and writes each declared output field into this workflow's Attributes by the same key -- the node that composes with Branch for real decisioning (docs/goals/0031-ai-node-family.md). Composition matches process-ai-completion: user content = Prompt + payload. Every declared field is required in the requested schema; a field the provider's response omits still appears in Attributes, zero-valued for its type.",
+		Description: "Sends a prompt (plus the running payload) to a Configure-authored AI provider, requests a JSON-schema-constrained structured response, and writes each declared output field into this workflow's Attributes by the same key -- the step that composes with Branch for real decisioning (docs/goals/0031-ai-node-family.md). Composition matches process-ai-completion: user content = Prompt + payload. Every declared field is required in the requested schema; a field the provider's response omits still appears in Attributes, zero-valued for its type.",
 		ConfigFields: []ConfigField{
 			{
 				Key: aiProviderIDConfigKey, Label: "AI provider",