Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.cyberdesk.io/llms.txt

Use this file to discover all available pages before exploring further.

When developers want to skip certain form fields in workflows, Cyberdesk provides the __EMPTY__ sentinel value. If you want to actively clear an already-filled field instead, use __CLEAR__. Cyberdesk also treats optional empty strings and null values as “not provided” when building run inputs, so they are normalized to __EMPTY__ before execution.

​
Simple Use Case

Problem: Sometimes workflows have optional form fields that should be skipped when no data is provided. Solution: Use __EMPTY__ to skip typing while maintaining workflow structure.

​
Example

# User provides username but no email
input_values = {"username": "john_doe", "email": "__EMPTY__"}

# Agent behavior:
type_text("john_doe")  # Types the username
type_text("__EMPTY__")  # Skips typing (no email provided)

​
Automatic __EMPTY__ for Missing Nested Fields

When using structured inputs with nested access, missing fields automatically become __EMPTY__: 
# Prompt uses: {patient.middle_name}
# Input only has first and last name:
input_values = {
    "patient": {
        "first_name": "John",
        "last_name": "Doe"
        # middle_name is not provided
    }
}

# Result: {patient.middle_name} → __EMPTY__
This allows workflows to gracefully handle optional nested fields without requiring explicit __EMPTY__ values.

​
Nested Access on __EMPTY__ Values

If a root variable is __EMPTY__ and you try to access a nested property on it, the result is also __EMPTY__: 
# Prompt uses: {extra_info.notes}
# User provides nothing for extra_info (auto-converted to __EMPTY__)
input_values = {"extra_info": "__EMPTY__"}

# Result: {extra_info.notes} → __EMPTY__
This allows workflows to gracefully skip entire optional sections without failing.
Type errors are different: If you try to access a nested property on a value that is a concrete type like a string or number (e.g., {patient.name.first} when patient.name is "John Doe"), the run fails immediately with a clear error. Only missing fields and __EMPTY__ values become __EMPTY__.

​
Input Schema Validation

When a workflow has an input schema, __EMPTY__ is treated as “not provided” for validation.
  • Object fields set to __EMPTY__ are ignored during schema validation.
  • Array entries set to __EMPTY__ are removed before validation.
  • This applies to auto-generated __EMPTY__ values too, including optional inputs left blank as empty strings or null.
This lets optional fields stay blank without failing validation, while required fields still need a real value unless your schema allows them to be omitted.

​
When to Use

  • Optional form fields that might not always be filled
  • Progressive workflows where some data is provided later
  • A/B testing different workflow variants
  • Graceful degradation when inputs are missing
  • Nested optional fields in structured input objects

​
Best Practice

# ✅ Good: Simple optional step
type_text("__EMPTY__", conditional="input_values['optional_field'] == '__EMPTY__'")

# ❌ Avoid: Complex logic - split workflows instead
This is a simple convenience feature for handling occasional missing data - not for complex workflow logic. If you need to blank out an existing value instead of skipping, see Clear Values.