workato-recipes

Workato Recipes Base Skill - Agent Instructions

This document provides foundational knowledge for AI agents to generate valid Workato recipe JSON. This is the base skill that connector-specific skills (stripe-recipes, salesforce-recipes, etc.) extend.

---

CRITICAL: Pre-Generation Checklist

For EXISTING projects (recipes already exist):

1. Read 1-2 existing .recipe.json files to understand project structure and patterns
2. Verify connection names - ask user for exact connection names in their workspace
3. Follow established folder structure - match where recipes and connections are organized

For GREENFIELD projects (no existing recipes):

1. Use skill templates as reference - see templates/ directory in each skill
2. Use standard file naming - {recipe_name}.recipe.json (lowercase, underscores for spaces)
3. Start recipe names with action verbs - "Create...", "Search...", "Update...", "Process..."
4. Ask user for connection names - exact names from their Workato workspace

ALWAYS (both scenarios):

1. Use descriptive UUIDs - {action}-{number} format (e.g., search-contact-001, return-success-005)
2. Use API endpoint triggers for testability (unless callable recipe is specifically needed)
3. Define all response codes upfront in the trigger (200, 400, 500 at minimum)

---

UUID Format (MANDATORY)

ALWAYS use descriptive UUIDs, regardless of what existing recipes in a project use:

search-contact-001
create-customer-002
if-found-003
return-success-004
return-error-005

NEVER use random hex UUIDs:

a1b2c3d4-e5f6-7890-abcd-ef1234567890  ← ANTI-PATTERN
11111111-1111-1111-1111-111111111111  ← ANTI-PATTERN

NOTE: Recipes created in Workato's declarative UI have random hash UUIDs. This is a platform limitation, NOT a pattern to follow. When you see random UUIDs in existing recipes, do NOT copy them. Always use descriptive UUIDs for new recipes and actions.

---

Recipe JSON Structure (CRITICAL)

The code field is an OBJECT (the trigger itself), NOT an array wrapped in a recipe object.

Key points:

• code is the trigger object directly, not wrapped in recipe
• code is NOT an array - actions go inside code.block
• Trigger number starts at 0, not 1
• Trigger as should be "trigger" for callable recipes

See: fundamentals/recipe-structure.md for full structure with examples.

---

Action Numbering (CRITICAL)

Every block must have a sequential number field:

Block	Number
Trigger	0
First action	1
Second action	2
...	...

Non-sequential numbers cause "out of sequence" errors that block recipe activation. When modifying recipes, always renumber all actions sequentially.

---

Built-in Providers

The workato provider is built-in and should NOT be in the recipe's config array. Only include config entries for external connectors requiring authentication (e.g., salesforce, stripe, gmail).

---

Connection Configuration (CRITICAL)

DO NOT put config blocks inside actions. Connections are defined ONLY in the top-level config array. Actions reference connections implicitly through the provider field.

WRONG (config inside action):

{
  "provider": "gmail",
  "name": "adhoc_http",
  "config": {
    "account_id": {"name": "My Gmail"}
  }
}

CORRECT (top-level only):

{
  "code": { ... },
  "config": [
    {
      "keyword": "application",
      "provider": "gmail",
      "account_id": {"name": "My Gmail"}
    }
  ]
}

---

Filename Convention

Recipe filenames must match the recipe's name field, converted to lowercase with spaces replaced by underscores. Workato normalizes filenames on pull, so mismatches cause filename changes.

Recipe Name	Correct Filename
"name": "Search Contact By Email"	search_contact_by_email.recipe.json
"name": "Create Stripe Customer"	create_stripe_customer.recipe.json
"name": "Handle Dialog Submit"	handle_dialog_submit.recipe.json

---

CRITICAL: REST Connector Action Name

WARNING: The rest provider MUST use make_request_v2 as its action name — NOT __adhoc_http_action. Using the wrong action name causes Workato to silently strip all input config on import. See patterns/adhoc-http-actions.md for the full make_request_v2 structure and examples.

---

Quick Reference

Topic	Documentation
Recipe Structure	fundamentals/recipe-structure.md
Config Section	fundamentals/config-section.md
Datapill Syntax	fundamentals/datapill-syntax.md
Variables & Lists	patterns/variables-and-lists.md
If/Else	control-flow/if-else.md
Try/Catch	control-flow/try-catch.md
Foreach Loops	control-flow/foreach.md
Repeat While Loops	control-flow/repeat-while.md
Stop Action	control-flow/stop.md
API Endpoint Trigger	triggers/api-endpoint.md
Callable Recipe Trigger	triggers/callable-recipe.md
Scheduler Trigger	triggers/scheduler.md
Messaging Topic Trigger	triggers/messaging-topic.md
Publish to Topic Action	triggers/messaging-topic.md
Adhoc HTTP Actions	patterns/adhoc-http-actions.md
JWT Bearer Auth	patterns/jwt-auth.md
API Platform Artifacts	patterns/api-platform-artifacts.md

---

Table of Contents

1. Trigger Types
2. Calling Other Recipes
3. Response Actions
4. Block Requirements
5. Extended Schemas
6. Formula Syntax

---

Trigger Types

Workato supports multiple trigger types. Choose based on how the recipe will be invoked.

API Endpoint Trigger (Recommended for Testing)

Use when: Recipe should be callable via external HTTP request (curl, webhooks, third-party systems).

Provider: workato_api_platform Action: receive_request

RECOMMENDATION: Use API endpoint triggers for most recipes. They're easier to test via curl and more practical for real integrations than callable recipes.

Note: API endpoint recipes require companion .api_endpoint.json and .api_group.json files in addition to the recipe JSON. See patterns/api-platform-artifacts.md for the complete artifact set and file formats.

Complete API Endpoint Example

{
  "number": 0,
  "provider": "workato_api_platform",
  "name": "receive_request",
  "as": "trigger",
  "keyword": "trigger",
  "input": {
    "request": {
      "content_type": "json",
      "schema": [
        {
          "name": "email",
          "label": "Email",
          "type": "string",
          "control_type": "text",
          "optional": false,
          "hint": "Customer email address"
        },
        {
          "name": "name",
          "label": "Name",
          "type": "string",
          "control_type": "text",
          "optional": false
        },
        {
          "name": "company",
          "label": "Company",
          "type": "string",
          "control_type": "text",
          "optional": true,
          "hint": "Optional company name"
        }
      ]
    },
    "response": {
      "content_type": "json",
      "responses": [
        {
          "name": "Success",
          "http_status_code": "200"
        },
        {
          "name": "Created",
          "http_status_code": "201"
        },
        {
          "name": "Bad Request",
          "http_status_code": "400"
        },
        {
          "name": "Server Error",
          "http_status_code": "500"
        }
      ]
    }
  },
  "extended_output_schema": [
    {
      "label": "Request",
      "name": "request",
      "type": "object",
      "properties": [
        {
          "name": "email",
          "label": "Email",
          "type": "string",
          "control_type": "text"
        },
        {
          "name": "name",
          "label": "Name",
          "type": "string",
          "control_type": "text"
        },
        {
          "name": "company",
          "label": "Company",
          "type": "string",
          "control_type": "text"
        }
      ]
    }
  ],
  "block": [
    // Actions go here
  ]
}

Request Schema Fields

Field	Required	Description
name	Yes	Field identifier (used in datapills)
label	Yes	Display label in UI
type	Yes	Data type: string, integer, boolean, date, date_time
control_type	Yes	UI control: text, number, checkbox, date, select
optional	Yes	true for optional, false for required
hint	No	Help text for the field

Multiple Response Codes

Define all possible HTTP responses in the trigger. The return_response action references these by name:

"responses": [
  { "name": "Success", "http_status_code": "200" },
  { "name": "Created", "http_status_code": "201" },
  { "name": "Bad Request", "http_status_code": "400" },
  { "name": "Conflict", "http_status_code": "409" },
  { "name": "Server Error", "http_status_code": "500" }
]

Datapill Paths for Request Fields

Access request fields directly (no body wrapper):

"path": ["request", "email"]
"path": ["request", "name"]
"path": ["request", "company"]

WRONG:

"path": ["request", "body", "email"]

Testing with curl

curl -X POST "https://apim.workato.com/your-workspace/your-endpoint" \
  -H "API-TOKEN: your-api-token" \
  -H "Content-Type: application/json" \
  -d '{"email": "[email protected]", "name": "Test User"}'

See: triggers/api-endpoint.md

Callable Recipe Trigger

Use when: Recipe should be called by other Workato recipes (internal).

Provider: workato_recipe_function Action: execute

{
  "provider": "workato_recipe_function",
  "name": "execute",
  "keyword": "trigger",
  "input": {
    "parameters_schema_json": "[...]",
    "result_schema_json": "[...]"
  }
}

See: triggers/callable-recipe.md

Scheduler Trigger

Use when: Recipe should run on a recurring time-based schedule (daily digests, periodic sync jobs) rather than in response to an external event.

Provider: clock Action: scheduled_event

See: triggers/scheduler.md

Choosing a Trigger Type

Scenario	Trigger Type
External API access needed	API Endpoint
Called by other recipes only	Callable Recipe
React to messages from other recipes	Messaging Topic (subscriber)
Receive external webhooks	Webhook
Time-based execution	Scheduler

---

Calling Other Recipes

When a recipe needs to call another callable recipe, use the workato_recipe_function provider with action type call.

CRITICAL: flow_id Requires zip_name

The flow_id object MUST include ALL fields, including zip_name.

Field	Required	Description
name	Yes	Recipe display name
folder	Yes	Folder containing the recipe
folder_full_path	Yes	Full path from Home
zip_name	YES	Path to recipe JSON file

CRITICAL WARNING: Missing zip_name causes RECIPE MUTATION AT RUNTIME. Without zip_name, Workato will unpredictably modify the recipe's metadata during execution. This corruption persists and breaks all future invocations. The recipe will appear valid during import/testing but will corrupt itself when actually invoked. This is worse than a silent failure - it permanently corrupts the recipe.

WRONG (missing zip_name - WILL CORRUPT THE RECIPE):

"flow_id": {
  "name": "Search contact by email",
  "folder": "atomic-salesforce-recipes",
  "folder_full_path": "Home/atomic-salesforce-recipes"
}

CORRECT (includes zip_name):

"flow_id": {
  "name": "Search contact by email",
  "folder": "atomic-salesforce-recipes",
  "folder_full_path": "Home/atomic-salesforce-recipes",
  "zip_name": "atomic-salesforce-recipes/search_contact_by_email.recipe.json"
}

Complete call action example:

{
  "provider": "workato_recipe_function",
  "name": "call",
  "as": "call_search_contact",
  "keyword": "action",
  "input": {
    "flow_id": {
      "name": "Search contact by email",
      "folder": "atomic-salesforce-recipes",
      "folder_full_path": "Home/atomic-salesforce-recipes",
      "zip_name": "atomic-salesforce-recipes/search_contact_by_email.recipe.json"
    },
    "parameters": {
      "email": "#{_dp('{\"pill_type\":\"output\",\"provider\":\"workato_api_platform\",\"line\":\"api_trigger\",\"path\":[\"request\",\"email\"]}')}"
    }
  }
}

---

Response Actions

Workato provides different actions for returning data based on the trigger type.

API Endpoint Response Action (return_response)

Use when: Recipe uses workato_api_platform trigger and needs to return HTTP response.

Provider: workato_api_platform Action: return_response

How pick_list Maps Response Names to HTTP Codes

The pick_list in extended_input_schema maps the response names (defined in trigger) to HTTP status codes:

"pick_list": [
  ["Success", "200"],      // "Success" from trigger → HTTP 200
  ["Created", "201"],      // "Created" from trigger → HTTP 201
  ["Bad Request", "400"],  // "Bad Request" from trigger → HTTP 400
  ["Server Error", "500"]  // "Server Error" from trigger → HTTP 500
]

CRITICAL: The first element (e.g., "Success") must match exactly the name field from the trigger's responses array.

Complete return_response Example

{
  "number": 5,
  "provider": "workato_api_platform",
  "name": "return_response",
  "as": "return_success",
  "keyword": "action",
  "uuid": "return-success-005",
  "input": {
    "http_status_code": "200",
    "response": {
      "customer_id": "#{_dp('{\"pill_type\":\"output\",\"provider\":\"stripe\",\"line\":\"create_customer\",\"path\":[\"body\",\"id\"]}')}",
      "success": "true",
      "error_message": "=null"
    }
  },
  "extended_input_schema": [
    {
      "change_on_blur": true,
      "control_type": "select",
      "extends_schema": true,
      "label": "Response",
      "name": "http_status_code",
      "pick_list": [
        ["Success", "200"],
        ["Created", "201"],
        ["Bad Request", "400"],
        ["Server Error", "500"]
      ],
      "type": "string"
    },
    {
      "label": "Response body",
      "name": "response",
      "type": "object",
      "properties": [
        {
          "control_type": "text",
          "label": "Customer ID",
          "name": "customer_id",
          "type": "string"
        },
        {
          "control_type": "checkbox",
          "label": "Success",
          "name": "success",
          "type": "boolean"
        },
        {
          "control_type": "text",
          "label": "Error Message",
          "name": "error_message",
          "type": "string"
        }
      ]
    }
  ],
  "extended_output_schema": [
    {
      "change_on_blur": true,
      "control_type": "select",
      "extends_schema": true,
      "label": "Response",
      "name": "http_status_code",
      "pick_list": [
        ["Success", "200"],
        ["Created", "201"],
        ["Bad Request", "400"],
        ["Server Error", "500"]
      ],
      "type": "string"
    },
    {
      "label": "Response body",
      "name": "response",
      "type": "object",
      "properties": [
        {
          "control_type": "text",
          "label": "Customer ID",
          "name": "customer_id",
          "type": "string"
        },
        {
          "control_type": "checkbox",
          "label": "Success",
          "name": "success",
          "type": "boolean"
        },
        {
          "control_type": "text",
          "label": "Error Message",
          "name": "error_message",
          "type": "string"
        }
      ]
    }
  ]
}

Multiple Return Actions Pattern

Use separate return_response actions for different scenarios:

return_success (200)     → Happy path
return_created (201)     → New record created
return_not_found (200)   → Search found nothing (still success)
return_bad_request (400) → Invalid input
return_error (500)       → Caught exception

Each action has the same extended_input_schema (with all response codes in pick_list) but different input.http_status_code values.

Legacy Format Reference

Older recipes may use this more verbose format with toggleCfg and toggle_field:

{
  "provider": "workato_api_platform",
  "name": "return_response",
  "as": "return_success",
  "keyword": "action",
  "toggleCfg": {
    "response.success": true
  },
  "input": {
    "http_status_code": "200",
    "response": {
      "id": "#{datapill}",
      "success": "true"
    }
  },
  "extended_output_schema": [
    {
      "change_on_blur": true,
      "control_type": "select",
      "extends_schema": true,
      "label": "Response",
      "name": "http_status_code",
      "pick_list": [
        ["Success", "200"],
        ["Error", "400"]
      ],
      "type": "string"
    },
    {
      "label": "Response body",
      "name": "response",
      "properties": [
        {
          "control_type": "text",
          "label": "ID",
          "name": "id",
          "type": "string"
        },
        {
          "control_type": "checkbox",
          "label": "Success",
          "render_input": "boolean_conversion",
          "parse_output": "boolean_conversion",
          "name": "success",
          "type": "boolean",
          "toggle_hint": "Select from option list",
          "toggle_field": {
            "label": "Success",
            "control_type": "text",
            "toggle_hint": "Use custom value",
            "name": "success",
            "type": "boolean"
          }
        }
      ],
      "type": "object"
    }
  ],
  "extended_input_schema": [
    {
      "change_on_blur": true,
      "control_type": "select",
      "extends_schema": true,
      "label": "Response",
      "name": "http_status_code",
      "pick_list": [
        ["Success", "200"],
        ["Error", "400"]
      ],
      "type": "string"
    },
    {
      "label": "Response body",
      "name": "response",
      "properties": [
        {
          "control_type": "text",
          "label": "ID",
          "name": "id",
          "type": "string"
        },
        {
          "control_type": "checkbox",
          "label": "Success",
          "render_input": "boolean_conversion",
          "parse_output": "boolean_conversion",
          "name": "success",
          "type": "boolean",
          "toggle_hint": "Select from option list",
          "toggle_field": {
            "label": "Success",
            "control_type": "text",
            "toggle_hint": "Use custom value",
            "name": "success",
            "type": "boolean"
          }
        }
      ],
      "type": "object"
    }
  ],
  "uuid": "return-success-001"
}

CRITICAL fields:

• http_status_code - HTTP status code (200, 400, 500, etc.)
• response - Response body matching the schema defined in trigger
• extended_input_schema.http_status_code.pick_list - Must map response names from trigger to status codes
• extended_input_schema.http_status_code.extends_schema: true - Required flag
• extended_input_schema.http_status_code.change_on_blur: true - Required flag
• extended_output_schema - Must mirror extended_input_schema
• toggleCfg - Optional, for boolean toggle fields

Common pattern:

• Success responses: http_status_code: "200" or "201"
• Error responses: http_status_code: "400", "409", or "500"
• The pick_list MUST match the response names defined in the trigger's response.responses array

Callable Recipe Response Action

Use when: Recipe uses workato_recipe_function trigger and needs to return data.

Provider: workato_recipe_function Action: return_result

{
  "provider": "workato_recipe_function",
  "name": "return_result",
  "as": "return_result",
  "keyword": "action",
  "input": {
    "result": {
      "id": "#{datapill}",
      "success": "true"
    }
  },
  "extended_output_schema": [
    {
      "label": "Result",
      "name": "result",
      "properties": [
        {
          "control_type": "text",
          "label": "ID",
          "name": "id",
          "type": "string"
        },
        {
          "control_type": "checkbox",
          "label": "Success",
          "name": "success",
          "type": "boolean"
        }
      ],
      "type": "object"
    }
  ],
  "extended_input_schema": [
    {
      "label": "Result",
      "name": "result",
      "properties": [
        {
          "control_type": "text",
          "label": "ID",
          "name": "id",
          "type": "string"
        },
        {
          "control_type": "checkbox",
          "label": "Success",
          "name": "success",
          "type": "boolean"
        }
      ],
      "type": "object"
    }
  ],
  "uuid": "return-result-001"
}

Key fields:

• result - Result object matching the schema defined in trigger's result_schema_json

---

Block Requirements

Every block in a recipe requires these fields:

Field	Required	Description
number	Yes	Sequential step number
keyword	Yes	Block type: trigger, action, if, else, try, catch, foreach, stop, repeat, while_condition
uuid	Yes	Unique identifier (max 36 chars)
as	Yes*	Step alias for datapill references

*Required for actions, triggers, catch, foreach, and repeat. Optional for try, if, and else blocks. NOT used on while_condition.

Provider Field Rules

Block Type	Provider
Trigger	Required (e.g., workato_api_platform)
Action	Required (e.g., stripe, salesforce)
If/Else	NO provider field
Foreach	NO provider field
Repeat / While Condition	NO provider field
Catch	"provider": null (explicitly null)

UUID Guidelines

• Must be unique within the recipe
• Max 36 characters (platform will reject recipe if exceeded)
• Use descriptive names: "search-customer-001", "return-success-001"

WARNING: UUIDs longer than 36 characters will cause the recipe to be rejected during import. Keep names concise.

---

Extended Schemas

Actions that return or accept complex data need extended schemas.

CRITICAL: Schema Completeness Requirement

WARNING: The extended_input_schema MUST fully define ALL fields in the action's input object, including nested objects. If any input field is missing from the schema, Workato will silently drop that data during execution, causing cascading failures.

This is a Workato platform behavior that affects ALL connectors. Complex connectors (Salesforce, NetSuite, etc.) are particularly vulnerable due to deeply nested input structures.

Symptoms of incomplete schemas:

• Input data silently dropped (no error, just missing)
• API calls missing required parameters
• Downstream steps fail due to missing data
• Difficult to debug because the recipe structure looks correct

Agent requirement: When generating recipes with custom actions or complex inputs, ALWAYS verify that extended_input_schema mirrors the complete input structure.

EXCEPTION — Native connector internal parameters: Native connector actions have built-in parameters handled internally by the connector. These must NEVER appear in extended_input_schema — if included, Workato creates duplicate fields in the UI, with the input value routing to the EIS copy (leaving the native field blank). This applies to ALL native connectors, not just Salesforce. Examples:

• Salesforce search_sobjects: sobject_name, limit are internals. Only user-facing filter fields (Id, AccountId, Email) go in EIS.
• Salesforce search_sobjects_soql: query is an internal. Empty EIS is correct.
• Jira search_issues_by_JQL: jql is the internal field name (NOT query). Empty EIS is correct. Note: action name is case-sensitive — must be uppercase JQL.
• General rule: If an action has a built-in required field visible in the UI, do NOT redeclare it in EIS. Use input.{field_name} to set its value; the field name must match the connector's internal name (pull a blank action from the UI to discover it). See the connector-specific skill files for details on which parameters are connector internals.

extended_output_schema

Defines the output fields available from an action:

"extended_output_schema": [
  {
    "label": "Customer ID",
    "name": "id",
    "type": "string",
    "control_type": "text"
  }
]

extended_input_schema

Defines the input fields for an action. Must mirror the input structure exactly.

Simple flat input:

"input": {
  "email": "[email protected]"
}

"extended_input_schema": [
  {
    "label": "Email",
    "name": "email",
    "type": "string",
    "control_type": "text",
    "optional": false
  }
]

Nested input (e.g., custom HTTP actions):

"input": {
  "path": "/v1/customers/search",
  "verb": "get",
  "input": {
    "schema": "[...]",
    "data": {
      "query": "email:'[email protected]'",
      "limit": "1"
    }
  }
}

"extended_input_schema": [
  {
    "label": "Path",
    "name": "path",
    "type": "string",
    "control_type": "text"
  },
  {
    "label": "Verb",
    "name": "verb",
    "type": "string",
    "control_type": "select"
  },
  {
    "label": "Request URL parameters",
    "name": "input",
    "type": "object",
    "properties": [
      {
        "label": "Schema",
        "name": "schema",
        "type": "string",
        "control_type": "text"
      },
      {
        "label": "Data",
        "name": "data",
        "type": "object",
        "properties": [
          {
            "label": "Query",
            "name": "query",
            "type": "string",
            "control_type": "text"
          },
          {
            "label": "Limit",
            "name": "limit",
            "type": "string",
            "control_type": "text"
          }
        ]
      }
    ]
  }
]

Schema Validation Checklist

Before finalizing any action block, verify:

• Every field in input has a corresponding entry in extended_input_schema
• Nested objects in input have matching nested properties in the schema
• Field names match exactly between input and schema
• For custom HTTP actions, the nested input.data object is fully defined

---

Formula Syntax

Workato formulas use a restricted subset of Ruby methods — not all Ruby methods are supported. Using an unsupported method will block recipe activation with no clear error message. Only use methods from the allowlist below.

Formula Prefix

Formulas are prefixed with =:

"input": {
  "field_name": "=formula_expression"
}

Conditional Field Updates (Ternary Operator)

Use .present? checks with ternary operator to conditionally set fields:

"FirstName": "=_dp('{\"pill_type\":\"output\",\"provider\":\"workato_recipe_function\",\"line\":\"trigger\",\"path\":[\"parameters\",\"first_name\"]}').present? ? _dp('{\"pill_type\":\"output\",\"provider\":\"workato_recipe_function\",\"line\":\"trigger\",\"path\":[\"parameters\",\"first_name\"]}') : skip"

Pattern:

=datapill.present? ? datapill : skip

• present? - Checks if value exists and is not empty
• ? value : skip - If present, use value; otherwise, skip the field
• skip - Special keyword to exclude field from the action

Supported Formula Methods (COMPLETE ALLOWLIST)

Workato formulas are an allowlist of Ruby methods. If a method is not listed here, do not use it — it will block recipe activation. This list is sourced from the Workato formula documentation.

String Methods

Method	Description
blank?	True if nil, empty, or whitespace only
present?	True if not blank
presence	Returns value if present, nil otherwise
include?	True if string contains substring
exclude?	True if string does not contain substring
match?	True if string matches regex pattern
starts_with?	True if string starts with prefix
ends_with?	True if string ends with suffix
is_true?	True if value is truthy
is_not_true?	True if value is falsy
strip	Remove leading/trailing whitespace
lstrip	Remove leading whitespace
rstrip	Remove trailing whitespace
upcase	Convert to uppercase
downcase	Convert to lowercase
capitalize	Capitalize first letter
titleize	Capitalize first letter of each word
reverse	Reverse the string
gsub	Replace all occurrences of pattern
sub	Replace first occurrence of pattern
strip_tags	Remove HTML tags
scrub	Replace invalid byte sequences
parameterize	Convert to URL-safe slug
quote	Wrap in quotes
length	Number of characters
slice	Extract substring by position
scan	Find all matches of pattern
split	Split into array by delimiter
ljust	Left-justify with padding
rjust	Right-justify with padding
encode	Encode to specified encoding
transliterate	Transliterate to ASCII
bytes	Convert to byte array
bytesize	Size in bytes
byteslice	Extract bytes by position
to_s	Convert to string
to_i	Convert to integer
to_f	Convert to float
ordinalize	Convert number to ordinal string (1st, 2nd, 3rd)
to_country_alpha2	Convert country name to ISO alpha-2 code
to_country_alpha3	Convert country name to ISO alpha-3 code
to_country_name	Convert country code to name
to_currency	Format as currency string
to_currency_code	Convert to currency code
to_currency_name	Convert to currency name
to_currency_symbol	Convert to currency symbol
to_phone	Format as phone number
to_state_code	Convert state name to code
to_state_name	Convert state code to name

Number Methods

Method	Description
abs	Absolute value
round	Round to specified precision
ceil	Round up
floor	Round down
even?	True if even
odd?	True if odd
blank?	True if nil
present?	True if not nil
presence	Returns value if present, nil otherwise
to_i	Convert to integer
to_f	Convert to float
to_s	Convert to string
to_currency	Format as currency
to_phone	Format as phone number

Date/Time Methods

Method	Description
now	Current timestamp
today	Current date
from_now	Duration from now (e.g., 30.days.from_now)
ago	Duration ago (e.g., 1.hour.ago)
strftime	Format with pattern (e.g., strftime('%Y-%m-%dT%H:%M:%SZ'))
in_time_zone	Convert to timezone (e.g., in_time_zone("UTC"))
beginning_of_hour	Start of current hour
beginning_of_day	Start of current day
beginning_of_week	Start of current week
beginning_of_month	Start of current month
beginning_of_year	Start of current year
end_of_month	End of current month
wday	Day of week (0=Sunday)
yday	Day of year
yweek	Week of year
dst?	True if daylight saving time
to_date	Convert to date
to_time	Convert to time
to_i	Convert to Unix epoch integer

Array/List Methods

Method	Description
first	First element
last	Last element
index	Position of element
count	Number of elements
length	Number of elements
where	Filter by condition
except	Exclude by condition
pluck	Extract field values
format_map	Format each element
join	Combine into string with separator
smart_join	Join, skipping blank values
concat	Append another array
reverse	Reverse order
sum	Sum of elements
uniq	Remove duplicates
flatten	Flatten nested arrays
max	Maximum value
min	Minimum value
compact	Remove nil values
blank?	True if empty
include?	True if contains element
exclude?	True if does not contain element
present?	True if not empty
presence	Returns array if present, nil otherwise
to_csv	Convert to CSV string
to_json	Convert to JSON string
to_xml	Convert to XML string
from_xml	Parse XML string
encode_www_form	URL-encode as form data
encode_url	URL-encode a string
to_param	Convert to URL parameter string
keys	Hash keys as array
values	Hash values as array

Any method not listed above will block recipe activation. Workato's formula language is a strict allowlist — standard Ruby methods like .each, .map, .chomp, .merge, .utc, and .to_a do not exist and will cause silent import failures.

Common equivalents:

• For UTC timestamps: use in_time_zone("UTC") or strftime('%Y-%m-%dT%H:%M:%SZ') (not .utc)
• For array access: use .first / .last (not [0] or [n])
• For JSON parsing: use the json_parser connector's parse_json_v2 action (not .parse_json — no such formula method exists). See adhoc-http-actions.md for the pattern.
• For filtering/mapping arrays: use .where, .pluck, .format_map (not .select, .map, .each)

Formula Examples

Conditional assignment:

"Email": "=_dp('{...email...}').present? ? _dp('{...email...}') : skip"

String transformation:

"Status": "=_dp('{...status...}').upcase"

Null assignment:

"field_name": "=null"

String Concatenation

When concatenating strings from multiple datapills, use formula mode (= prefix) with Ruby + operator.

WRONG (mixing #{} interpolation with concatenation - INVALID):

"guest_name": "#{_dp('{...first_name...}')} + ' ' + _dp('{...last_name...}'}"

CORRECT (use = prefix, no #{} wrapper):

"guest_name": "=_dp('{\"pill_type\":\"output\",\"provider\":\"workato_recipe_function\",\"line\":\"trigger\",\"path\":[\"parameters\",\"first_name\"]}') + ' ' + _dp('{\"pill_type\":\"output\",\"provider\":\"workato_recipe_function\",\"line\":\"trigger\",\"path\":[\"parameters\",\"last_name\"]}')"

Pattern:

"field": "=_dp('{...pill1...}') + ' ' + _dp('{...pill2...}')"

Key rules:

• Use = prefix (formula mode) for concatenation
• Do NOT wrap with #{}
• Use Ruby + operator between strings
• Literal strings must be quoted: ' ' or '-'

Conditional Defaults

Provide a default value when an optional parameter is missing using .present? with ternary operator:

Pattern:

"field": "=_dp('{...}').present? ? _dp('{...}') : 'default_value'"

Example (default status to 'Reserved' when not provided):

"Status__c": "=_dp('{\"pill_type\":\"output\",\"provider\":\"workato_recipe_function\",\"line\":\"trigger\",\"path\":[\"parameters\",\"status\"]}').present? ? _dp('{\"pill_type\":\"output\",\"provider\":\"workato_recipe_function\",\"line\":\"trigger\",\"path\":[\"parameters\",\"status\"]}') : 'Reserved'"

Key points:

• .present? checks if value exists and is not empty
• The datapill must appear twice (once for check, once for value)
• Default value must be quoted if it's a string

Error Handling Return Values (CRITICAL)

When recipes have required return parameters, ALL code paths (success AND catch blocks) must provide values. In catch blocks where data isn't available, use =null:

WRONG: "customer_id": "" CORRECT: "customer_id": "=null"

Selecting Between Multiple Sources (Ternary)

When you need to return a value from one of multiple possible sources (e.g., search result OR create result), use ternary syntax:

"customer_id": "=_dp('{...search_result...}').present? ? _dp('{...search_result...}') : _dp('{...create_result...}')"

This avoids the need for intermediate variables and works at both validation and runtime.

---

Connection Configuration

Same-Folder References

When connections are in the same folder as recipes, use empty string for folder:

"account_id": {
  "zip_name": "my_connection.connection.json",
  "name": "My Connection Name",
  "folder": ""
}

Different-Folder References

When connections are in a different folder:

"account_id": {
  "zip_name": "Connections/my_connection.connection.json",
  "name": "My Connection Name",
  "folder": "Connections"
}

---

Common Mistakes

Mistake	Symptom	Fix
recipe.code[] wrapper	Recipe doesn't render	Use code as object directly
Non-sequential action numbers	Activation error	Renumber sequentially from 0
workato provider in config	Push error	Remove - it's built-in
Empty string for required params	Activation error	Use =null
+ concat with datapills at import	Validation error	Use ternary or single datapill
Random hex UUIDs	Poor maintainability	Always use descriptive UUIDs (don't copy existing random UUIDs)
Copying patterns from declarative UI recipes	Various errors	Use skill templates, not UI-generated recipes as reference

---

---

Validation

See validation-checklist.md for the consolidated recipe validation checklist.

---

References

• Fundamentals: See fundamentals/ directory for recipe structure, config, and datapill syntax
• Triggers: See triggers/ directory for detailed trigger documentation
• Control Flow: See control-flow/ directory for if/else, try/catch, foreach, stop, and repeat-while patterns
• Templates: See templates/ directory for starter templates