# YY Workload Receipt

- Run: `example-openai-usage`
- Input fingerprint: `31b453e97cb4913088bc49ae39603028d2294db3de2900bff1429b44bf412e32`
- Valid events: 2

## Observed workload

- Model requests: 2
- Tokens: 150 input, 50 output, 10 cached, 7 reasoning
- Token total (input + output): 200
- Observed event duration: 0 ms
- Tool calls: 0 (0 ok, 0 error, 0 timeout)
- Retries: 0
- Error events: 0
- Approval observations: 0
- Observed approval wait: not observed

## Model breakdown

- `openai/example-model`: 2 requests, 200 input+output tokens, 0 ms observed

## Evidence gaps

- `MODEL_DURATION_NOT_OBSERVED`: 2 at [1, 2]

## Boundaries

This receipt reports observed metadata only. It does not retain prompts, messages, tool arguments, tool results, or error messages. It is not a cost estimate, security audit, compliance certification, or hardware recommendation.