Secret in agent context
Judges: text about to be sent to model. Questions: noul:secret - holds credential?. Action: Redact before call.
AA clean typed decision (yes/no, a pick from a list, or a level on a scale) that is concrete, repeats routinely, and has meaningful value.5.31
Key facts
- Vertical
- Cross-industry
- Function
- Agents & dev
- Status
- Idea
- Volume
- routine
- Value
- major
- Risk
- high
- Evidence
- —
- Flags
- human-review
Build this with a classifier
Define a typed decision with a bounded answer, then evaluate it on examples.
{
"decision_type": "yes_no",
"question": "Does this input match the decision in “Secret in agent context”?",
"input": "<input to classify>",
"output": "yes | no"
}Related use cases
Cite this
Copy a link in your preferred format.