Cognotik | GenerateImage Task

⚙️ TaskConfig.json

{
  "files": ["assets/hero.png"],
  "task_description": "A futuristic city 
    at sunset, synthwave style, 
    highly detailed, 8k",
  "related_files": ["docs/style.md"]
}

→

👁️ Session Preview

Generating Image: assets/hero.png

🖼️ [ Synthwave Cityscape Preview ]

✔ Successfully generated and saved image.

📂 Test Workspace: ImageGenerationTask

Live view of the webui/workspaces/ImageGenerationTask/ directory containing generated assets and logs.

Execution Configuration

Field	Type	Description
`files` *	`List<String>`	The image file to be created. Must be a single file ending in `.png`, `.jpg`, or `.jpeg`.
`task_description`	`String`	Detailed description of the image including subject, style, composition, and mood.
`related_files`	`List<String>`	Reference images or style guides to provide visual context to the generator.
`task_dependencies`	`List<String>`	List of task IDs that must complete before this generation starts.

Kotlin Usage

Invoke via UnifiedHarness for headless automation:

import com.simiacryptus.cognotik.plan.tools.file.ImageGenerationTask.Companion.GenerateImage
import com.simiacryptus.cognotik.plan.tools.file.ImageGenerationTask.GenerateImageTaskExecutionConfigData

harness.runTask(
    taskType = GenerateImage,
    typeConfig = TaskTypeConfig(),
    executionConfig = GenerateImageTaskExecutionConfigData(
        files = listOf("output/branding/logo_v1.png"),
        task_description = "Minimalist tech logo, vector style, blue and silver",
        related_files = listOf("docs/brand_guidelines.pdf")
    ),
    workspace = File("./project-assets"),
    autoFix = true
)

Test Case Example

Example of a self-healing UI asset generation loop:

// 1. Generate the asset
// 2. Run a visual regression test
// 3. If test fails, the agent uses the error log as context to regenerate
val config = GenerateImageTaskExecutionConfigData(
    files = listOf("src/main/resources/icon.png"),
    task_description = "App icon, 512x512, flat design",
    task_dependencies = listOf("ui_spec_analysis")
)

Prompt Segment

The following instruction is provided to the orchestrator:

GenerateImage - Create high-quality images using AI generation models.
  * Specify a single output file path (png, jpg, or jpeg).
  * Provide a detailed task_description covering style, composition, and mood.
  * Use related_files to provide visual context or style references.
  * Useful for UI mockups, illustrations, and visual assets.

Execution Configuration

Task Lifecycle

Kotlin Usage

Test Case Example

Prompt Segment