Test fixture
Bug fixes, edge conditions, and minimal patch accuracy.
The model receives the prompt (and optional system message). The run uses scorer rubric_json_metrics with the JSON configuration below. Pass/fail and partial credit are determined entirely by that scorer against the model output; no human grading.
Return JSON only with keys diagnosis, fix, tests. A config object is directly assigned to a new variable and then modified; all callers that hold the original reference see unexpected mutations. Identify the bug and fix it.
{
"metrics": {
"repro": {
"checks": [
{
"contains": [
"assignment"
]
},
{
"contains": [
"reference"
]
},
{
"contains": [
"config"
]
},
{
"contains": [
"mutation"
]
}
]
},
"hidden": {
"checks": [
{
"contains": [
"Object.assign"
]
},
{
"contains": [
"spread"
]
},
{
"contains": [
"copy"
]
}
]
},
"diagnose": {
"checks": [
{
"contains": [
"shared reference"
]
},
{
"contains": [
"same object"
]
},
{
"contains": [
"not a clone"
]
}
]
}
}
}temperature
0
max_tokens
420
timeout (s)
120
type
scored
file
debug-object-reference-v2.json