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 API client sends authentication tokens over plain HTTP instead of HTTPS, exposing them in transit.
{
"metrics": {
"correct": {
"checks": [
{
"contains": [
"plaintext"
]
},
{
"contains": [
"HTTPS"
]
},
{
"contains": [
"TLS"
]
},
{
"contains": [
"intercept"
]
},
{
"contains": [
"token"
]
}
]
},
"hidden": {
"checks": [
{
"contains": [
"HSTS"
]
},
{
"contains": [
"certificate"
]
},
{
"contains": [
"man-in-the-middle"
]
},
{
"contains": [
"redirect HTTP"
]
},
{
"contains": [
"TLS 1.2"
]
}
]
}
}
}temperature
0
max_tokens
420
timeout (s)
120
type
scored
file
sec-insecure-http-transport.json