Test fixture
Secure code changes, vulnerability recognition, and safe defaults.
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 verdict, risk, mitigation. An XML parser processes user uploaded XML with external entities enabled.
{
"metrics": {
"correct": {
"checks": [
{
"contains": [
"XXE"
]
},
{
"contains": [
"external entities"
]
},
{
"contains": [
"XML"
]
}
]
},
"hidden": {
"checks": [
{
"contains": [
"disable DTD"
]
},
{
"contains": [
"external entity"
]
},
{
"contains": [
"file disclosure"
]
}
]
}
}
}temperature
0
max_tokens
460
timeout (s)
120
type
scored
file
sec-xml-external-entity.json