JournalismReasoningTask

Execution Configuration

* Required fields

Task Execution Flow

1. Fact Verification: Identifies 5-10 key claims and checks against evidence.
2. Perspective Identification: Maps 4-6 stakeholder viewpoints and credibility.
3. Context Analysis: Researches historical background and broader implications.
4. Bias Detection: Analyzes framing issues and missing voices.
5. Angle Exploration: Generates creative approaches with newsworthiness scores.
6. Gap Identification: Finds unanswered questions and research approaches.
7. Editorial Synthesis: Final authoritative summary of findings.

The task produces a comprehensive Markdown report saved to the workspace as journalism_analysis_[timestamp].md.

Direct Task Invocation

val task = JournalismReasoningTask(
    orchestrationConfig = config,
    planTask = JournalismReasoningTaskExecutionConfigData(
        story_topic = "The rise of vertical farming",
        verify_facts = true,
        alternative_angles = 5
    )
)

// 1. Initialize the Harness
val harness = UnifiedHarness(serverless = true, openBrowser = false, ...)
harness.start()

// 2. Define the Job
val config = JournalismReasoningTaskExecutionConfigData(
    story_topic = "Impact of remote work on urban planning",
    input_files = listOf("docs/city_reports/*.pdf"),
    find_gaps = true
)

// 3. Execute
harness.runTask(
    taskType = JournalismReasoningTask.JournalismReasoning,
    typeConfig = TaskTypeConfig(),
    executionConfig = config,
    workspace = File("./investigation-output")
)

Prompt Segment

The following logic is injected into the LLM context:

JournalismReasoning - Investigate stories through journalistic principles
** Specify the story topic or event to investigate
** Define journalism elements: who, what, when, where, why, how
** Enable fact verification and source checking
** Identify multiple perspectives and stakeholder voices
** Detect potential biases and conflicts of interest