Use this skill when users need to create Custom Lightning Types (CLTs) for Einstein Agent actions or structured input/output schemas. Trigger when users mention CLT, Custom Lightning Types, JSON schemas for agents, type definitions, lightning__objectType, or editor/renderer configurations. For widget renditions that combine a CLT with a Widget bundle, use the platform-lightning-type-widget-coordinate orchestrator instead. This is complex - always use this skill for CLT work.
Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Une commande directe contourne le prompt de vérification. Examinez la source avant de l'exécuter.
Use this skill when users need to create Custom Lightning Types (CLTs) for Einstein Agent actions or structured input/output schemas. Trigger when users mention CLT, Custom Lightning Types, JSON schemas for agents, type definitions, lightning__objectType, or editor/renderer configurations. For widget renditions that combine a CLT with a Widget bundle, use the platform-lightning-type-widget-coordinate orchestrator instead. This is complex - always use this skill for CLT work.
Create Custom Lightning Types (CLTs) for structured inputs/outputs
Generate JSON Schema-based type definitions for Lightning Platform
Configure CLTs for Einstein Agent actions
Set up editor and renderer configurations for custom UI
Troubleshoot deployment errors related to Custom Lightning Types
Specification
CustomLightningType Metadata Specification
Overview & Purpose
Custom Lightning Types (CLTs) are JSON Schema-based type definitions used by the Lightning Platform (including Einstein Agent actions) to describe structured inputs/outputs and drive editor/renderer experiences.
Configuration
Choose referenced CLT pattern for nested objects - When you need a reusable or separately deployed nested type, create a CLT for that shape and reference it with "lightning:type": "c__<CLTName>". That string is the referenced type’s lightning:type value / FQN / registered identifier — not the JSON Schema title.
Choose standard Lightning types when the structure is simple and can be expressed with properties and supported primitive lightning:type identifiers.
Choose Apex class types (@apexClassType/...) when the structure already exists server-side and you want the Apex class to define the shape.
Include editor/renderer config only when you need custom UI behavior (custom LWC input/output components). Otherwise, omit.
Critical Rules (Read First)
CRITICAL: NEVER include the "$schema" field in schema.json
Salesforce CLT validator WILL REJECT schemas with this field, even if it's a valid JSON Schema $schema declaration.
Root object schemas MUST include:
"type": "object"
"title"
"lightning:type": "lightning__objectType"
"unevaluatedProperties": false
"unevaluatedProperties" is enforced as false by the CLT metaschema. Do not set it to true.
Root object schemas MUST NOT include"examples" when "unevaluatedProperties": false is set.
Nested objects (inside properties) MUST NOT set"lightning:type": "lightning__objectType".
Nested objects can be: references to other CLTs using c__<CLTName> syntax.
List/array properties are highly restricted by the CLT metaschema:
CRITICAL LIMITATION: the CLT metaschema may reject the items keyword entirely. Treat items as disallowed by default.
Root-level arrays (direct children of the root properties):
MUST include"lightning:type": "lightning__listType"
MUST NOT include"items"
OPTIONAL"type": "array"
Nested arrays (arrays inside nested objects) are the most common failure:
MUST include"type": "array"
MUST NOT include"lightning:type": "lightning__listType"
MUST NOT include"items"
When "unevaluatedProperties": false is set, any unknown keyword will fail validation. Prefer removing keywords over relaxing strictness.
Apex class CLTs are minimal:
Additional CLT Metaschema Validations
Org namespace validation: titles/descriptions and other string fields may be validated to ensure you are not using an org namespace in places that are disallowed.
Lightning type validation: CLTs are validated to prevent referencing internal namespaces (for example, disallowing types from internal namespaces like sfdc_cms where not permitted).
Object type validation: the CLT root is validated to ensure lightning:type is exactly lightning__objectType.
Primitive Types & Constraints
When you need the full list of supported primitive lightning:type identifiers, their constraints, and the allowed property-level keywords, read assets/primitive-types-and-constraints.md in this skill's directory.
Generation Workflow
Confirm the CLT approach
If referencing Apex: capture the exact class reference (@apexClassType/namespace__ClassName$InnerClass).
If using standard primitives: list the fields, their Lightning primitive types, and which fields are required.
Draft schema.json
DO NOT include "$schema" at the top
Start with the root object structure (required root fields).
Add properties using valid primitive lightning:type identifiers.
For nested-object properties, use CLT Reference pattern:
"lightning:type": "c__<CLTName>" to reference another CLT
The referenced CLT must be deployed to the org before the parent CLT.
For Apex-based nested objects: Use @apexClassType/... when structure exists server-side.
If the prompt explicitly requires true nested object output, prefer an Apex-based CLT (@apexClassType/...) for deploy-safe nested structures.
For arrays: follow the strict list rules (avoid items; avoid lightning:type on nested arrays).
Before deployment, verify exact lightning:type spellings (for example, use lightning__richTextType, not misspelled variants).
(Optional) Draft editor.json (only if custom UI is required)
Supported shape: Top-level editor object with editor.componentOverrides and editor.layout.
Top-level editor object.
Use editor.componentOverrides for component overrides.
Use editor.layout for layout.
DEPRECATED: Do NOT use propertyRenderers or view — these are legacy keys. Always use componentOverrides and layout instead.
Root override pattern (most common for fully custom editing UI):
Org/channel validation rejects nested object typing in LightningTypeBundle
Use CLT reference (c__<CLTName>) or Apex class types
Invalid CLT reference
Referenced CLT doesn't exist in org or incorrect syntax
Deploy the referenced CLT first; c__<CLTName> must match the referenced type’s lightning:type value / FQN / registered identifier, not title
Invalid or misspelled lightning:type (for example, lightning__richtextType instead of lightning__richTextType)
Incorrect generated type name
Cross-check all lightning:type values against supported type names and correct them before deployment
Array property rejected
Use of items (or lightning:type in nested arrays) rejected by validator
For nested arrays: keep only type: "array". For root arrays: use minimal structure; remove items if rejected
Apex-based CLT rejected
Extra fields added (e.g., type, properties)
Use only title, optional description, and lightning:type
Editor config rejected
Use of invalid patterns (es_property_editors/inputList, itemSchema) or unrecognized top-level keys
Use editor.componentOverrides and editor.layout; keep config minimal
additionalProperties error on layout attributes
Adding label or other attributes to lightning/propertyLayout
Only use property attribute in . Remove , , or any other attributes
Verification Checklist
Root schema has type: "object", title, lightning:type: "lightning__objectType", and unevaluatedProperties: false
Root schema does not include examples when strict validation is enabled
No nested object includes lightning:type: "lightning__objectType"
Arrays are defined minimally (especially nested arrays)
Only supported primitive lightning:type identifiers are used for leaf properties
Apex class CLTs contain only title/description and lightning:type: "@apexClassType/..."
Bundle structure and filenames match Lightning Types requirements
Editor config uses only allowed patterns (no es_property_editors/inputList, no itemSchema); use valid components (e.g. es_property_editors/inputText, es_property_editors/inputNumber) or custom c/ components
Renderer config uses output-style components (e.g. es_property_editors/outputText, es_property_editors/outputNumber) where applicable, not input editors
Layout configurations use lightning/propertyLayout with ONLY the property attribute (no label, title, or other attributes)
All attribute mappings ({!$attrs.propertyName}) reference properties that exist in the type schema
Custom LWC components have correct targets in -meta.xml: lightning__AgentforceInput for editors, lightning__AgentforceOutput for renderers
Root schema does NOT include field
Include onlytitle, description (optional), and lightning:type set to @apexClassType/....
Do not add type, properties, required, or unevaluatedProperties.
Custom LWC renderers/editors on an Apex class CLT MUST NOT use attributes in the root override — this overrides any prompt wording to the contrary. Since the schema has no properties block, there is nothing for {!$attrs.<name>} to resolve against — unevaluatedProperties: false will reject any attribute key (e.g. "You can't add the flightId property ... because the unevaluatedProperties keyword value is set to false"). Use "componentOverrides": { "$": { "definition": "c/<yourComponent>" } } with no attributes key at all. If the user's prompt explicitly asks for attribute mappings to specific fields (e.g. "with attribute mappings for fieldA, fieldB") on an Apex-class CLT renderer/editor, do NOT comply literally — omit attributes from the root override anyway, and say so in your response (e.g. "Note: attribute mappings were omitted because the backing type is an Apex-class CLT, which has no properties block to bind against").
No shell metacharacters that trigger the Vibes safe-shell filter. In any Bash tool call emitted by this skill, do NOT use command substitution ($(…) or backticks), process substitution (<(…), >(…)), brace expansion ({a,b,c} or {1..N}), or eval / exec. Vibes forces manual approval on these patterns even under Bypass mode and stalls the eval. Emit separate commands (mkdir -p a && mkdir -p b) or print each value with its own command and reason about the output rather than capturing it in a shell variable.
When passing schema data into a custom LWC, use attribute mapping with the {!$attrs.<name>} syntax: e.g. "attributes": { "myField": "{!$attrs.value}" } so the runtime binds schema values to your component's attributes.
CRITICAL: The <name> in {!$attrs.<name>} must be a property defined in your type schema. For example, if your schema has a property called temperature, use {!$attrs.temperature}, not {!$attrs.value} unless value is an actual property.
Valid editor components (examples): es_property_editors/inputText, es_property_editors/inputNumber, es_property_editors/inputRichText, es_property_editors/inputImage, es_property_editors/inputTextarea. Do not usees_property_editors/inputList.
Collection editor (for root-level lightning__listType properties): Use a collection-level override so the list is edited by a custom component: collection.editor.componentOverrides["$"] = { "definition": "c/<yourCollectionEditorComponent>" }. Alternatively, use editor.layout with lightning/propertyLayout and attributes.property = "<listPropertyName>" for default list editing.
editor.layout.children[*].definition = "lightning/propertyLayout" with attributes.property = "<propertyName>"
CRITICAL: lightning/propertyLayout only accepts the property attribute. Do NOT add label, title, or any other attributes — these will fail validation with additionalProperties: false errors.
Avoid known-invalid patterns:
Do not use es_property_editors/inputList.
Do not use itemSchema attributes.
(Optional) Draft renderer.json (only if custom UI or widget rendition is required)
Supported shape: Top-level renderer object with renderer.componentOverrides and renderer.layout.
Top-level renderer object.
Use renderer.componentOverrides for component overrides.
Use renderer.layout for layout.
DEPRECATED: Do NOT use propertyRenderers or view — these are legacy keys. Always use componentOverrides and layout instead.
Widget rendition pattern (reference an existing WidgetBundle as the root renderer): the renderer file is a thin wrapper that points at the widget by developer name ("definition": "@widget/c/<widgetDeveloperName>") and maps CLT schema properties to widget attributes via {!$attrs.<schemaPropertyName>}. Do NOT duplicate the widget body inside renderer.json. See references/widget-rendition.md for the full shape, binding rules, and constraints. For the full Apex → Lightning Type → Widget pipeline, use the platform-lightning-type-widget-coordinate orchestrator instead of this skill.
Root override pattern (most common for fully custom rendering UI with a custom LWC):
Use {!$attrs.<name>} in attribute mappings when binding schema data to custom renderer component attributes.
CRITICAL: Attribute mappings like {!$attrs.propertyName} must reference properties that actually exist in your type schema. Referencing non-existent properties will fail validation.
Type matching: Attribute values must match the expected type for the component. For example, if a component expects a string attribute, passing an integer will fail validation.
renderer.layout.children[*].definition = "lightning/propertyLayout" with attributes.property = "<propertyName>"
CRITICAL: Same as editor layouts, lightning/propertyLayout only accepts the property attribute. Do NOT add label, title, or any other attributes.
Collection renderer (for root-level lightning__listType properties): Use collection.renderer.componentOverrides["$"] = { "definition": "c/<yourListRendererComponent>" } or es_property_editors/genericListTypeRenderer to render the list.
(Optional) lightningTypes/<TypeName>/lightningDesktopGenAi/renderer.json
For Gen AI / Copilot the standard path is lightningDesktopGenAi/. Other targets (e.g. Experience Builder, Mobile Copilot, Enhanced Web Chat) use different subfolders when supported: experienceBuilder/, lightningMobileGenAi/, enhancedWebChat/.
(Optional - for widget rendition only) lightningTypes/<TypeName>/renderer.json
Configure custom LWC components (if using custom components)
CRITICAL: Custom LWC components referenced in editor/renderer configs MUST have the correct target configuration in their -meta.xml files:
For editor components (c/<componentName> used in editor.json): The LWC's -meta.xml file must include <target>lightning__AgentforceInput</target>
For renderer components (c/<componentName> used in renderer.json): The LWC's -meta.xml file must include <target>lightning__AgentforceOutput</target>
Without the correct target, deployment will fail with: Invalid target configuration. To use 'c/componentName' as a renderer/editor, your js-meta.xml file must include valid target 'lightning__AgentforceOutput/Input'.
