Interface: FormalJudgeRequest

@kortexya/reasoninglayer

---

@kortexya/reasoninglayer / Oversight / FormalJudgeRequest

Interface: FormalJudgeRequest

Defined in: src/types/oversight.ts:42

Request to run the FormalJudge oversight pipeline.

Properties

config?

optional config: JudgeConfigDto | null

Defined in: src/types/oversight.ts:52

Optional pipeline configuration overrides.

---

executionLog?

optional executionLog: string

Defined in: src/types/oversight.ts:50

Raw execution log (for full provenance).

---

finalOutput?

optional finalOutput: string

Defined in: src/types/oversight.ts:48

Agent’s final output/answer.

---

riskCategories?

optional riskCategories: string[]

Defined in: src/types/oversight.ts:59

Risk categories from the benchmark dataset (e.g., ["Lead to property loss", "Violate laws"]). Used by the 3-agent per-task decomposition pipeline to generate task-specific atomic conditions.

---

trajectory

trajectory: TrajectoryStepDto[]

Defined in: src/types/oversight.ts:46

Agent execution trajectory (ordered list of tool calls + results).

---

userIntent

userIntent: string

Defined in: src/types/oversight.ts:44

User instruction (what the agent was asked to do).