glassbox-ai/examples/code-generation.yml

Enterprise-grade AI testing framework with reliability, observability, and comprehensive validation

tests:
  - name: fibonacci_function
    description: Ensure clean, correct implementation of Fibonacci in Python
    prompt: "Write a Python function to compute the nth Fibonacci number."
    expect_contains: ["def", "return", "fibonacci"]
    expect_not_contains: ["eval", "import os"]
    max_tokens: 200
    max_cost_usd: 0.005
    block_patterns: []

  - name: hello_world_flask
    description: Validate it generates a minimal Flask app
    prompt: "Generate a Python Flask app that returns Hello World at the root route."
    expect_contains: ["Flask", "route", "Hello"]
    expect_not_contains: ["request.files", "sys.exit"]
    max_tokens: 250
    max_cost_usd: 0.007
    block_patterns: []

  - name: list_comprehension
    description: Generate Python list comprehension for squaring numbers
    prompt: "Show me a Python list comprehension to square the numbers from 1 to 10."
    expect_contains: ["[", "for", "in", "**2"]
    expect_not_contains: ["while", "lambda"]
    max_tokens: 100
    max_cost_usd: 0.003
    block_patterns: []