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 recursive flattenObject function throws on null even though null should be treated as a leaf value. Diagnose the base-case bug and fix it.
{
"metrics": {
"repro": {
"checks": [
{
"contains": [
"null"
]
},
{
"contains": [
"leaf"
]
},
{
"contains": [
"throws"
]
},
{
"contains": [
"base case"
]
}
]
},
"hidden": {
"checks": [
{
"contains": [
"typeof null"
]
},
{
"contains": [
"object"
]
},
{
"contains": [
"guard"
]
}
]
},
"diagnose": {
"checks": [
{
"contains": [
"recursion"
]
},
{
"contains": [
"flatten"
]
},
{
"contains": [
"runtime"
]
}
]
}
}
}temperature
0
max_tokens
420
timeout (s)
120
type
scored
file
debug-recursion-base-case-v2.json