id: edge-03-timeout-handling
name: "Edge Case 03: Long Running Task Handling"
description: |
  Tests that the agent handles potentially long-running tasks appropriately.
  
  The agent should:
  1. Recognize this could take time
  2. Provide progress updates or warnings
  3. Complete within reasonable timeout
  
  Validates:
  - Agent handles multi-step tasks
  - Agent provides appropriate feedback
  - Timeout handling works correctly

category: edge-case

prompts:
  - text: |
      List all TypeScript files in the evals/framework/src directory and count them.
      Then summarize what types of files are there.

approvalStrategy:
  type: auto-approve

behavior:
  mustUseAnyOf:
    - [glob]
    - [list]
    - [bash]
  minToolCalls: 1
  maxToolCalls: 5

expectedViolations:
  - rule: approval-gate
    shouldViolate: false
    severity: error
  - rule: tool-usage
    shouldViolate: false
    severity: warning

timeout: 90000

tags:
  - edge-case
  - timeout
  - multi-step
  - safe