id: golden-06-task-breakdown
name: "Golden 06: Task Breakdown - Agent Structures Complex Work"
description: |
  Tests that the agent can break down a complex task into steps.
  
  A quality agent should:
  1. Recognize this requires multiple steps
  2. Read relevant context/standards first
  3. Propose a structured plan
  
  This test asks the agent to read project standards before
  implementing something - a common pattern for quality work.
  
  Validates:
  - Agent reads context before proposing implementation
  - Agent structures its approach
  - context-loading evaluator works correctly
category: developer

prompts:
  - text: |
      Read the coding standards for this project from .opencode/context/core/standards/code.md and summarize the key points.

approvalStrategy:
  type: auto-approve

behavior:
  mustUseTools:
    - read
  minToolCalls: 1
  requiresContext: true

expectedViolations:
  - rule: context-loading
    shouldViolate: false
    severity: error
  - rule: approval-gate
    shouldViolate: false
    severity: error

timeout: 60000

tags:
  - golden
  - task-breakdown
  - context-awareness
  - safe