Skip to content

Interface: FormalJudgeResponse

@kortexya/reasoninglayer


@kortexya/reasoninglayer / FormalJudgeResponse

Defined in: src/types/oversight.ts:195

Response from the FormalJudge oversight pipeline.

optional agentSubVerdicts?: AgentSubVerdictDto[]

Defined in: src/types/oversight.ts:197

Per-agent sub-verdicts for multi-agent trajectories.


optional certificate?: CertificateDto | null

Defined in: src/types/oversight.ts:199

Full validity certificate with per-layer attestations (if safe and enabled).


optional certificateHash?: string | null

Defined in: src/types/oversight.ts:201

Merkle root hash of the validity certificate (if safe).


optional deceptionFactMap?: object | null

Defined in: src/types/oversight.ts:203

Deception fact map: all 25 atomic deception facts with boolean answers.


optional detectedDeceptionPredicates?: string[]

Defined in: src/types/oversight.ts:205

Deception predicates that triggered.


elapsedMs: number

Defined in: src/types/oversight.ts:207

Processing time in milliseconds.


extractionStats: ExtractionStatsDto

Defined in: src/types/oversight.ts:209

Extraction statistics (deterministic vs. semantic).


fabricationsDetected: number

Defined in: src/types/oversight.ts:211

Number of fabricated facts detected.


optional faithfulnessScore?: number | null

Defined in: src/types/oversight.ts:213

Faithfulness/deception detection score (if enabled).


fixSuggestions: FixSuggestionDto[]

Defined in: src/types/oversight.ts:215

Actionable fix suggestions for violations.


optional isDeceptive?: boolean | null

Defined in: src/types/oversight.ts:217

Whether the master deception predicate triggered.


layerResults: LayerResultDto[]

Defined in: src/types/oversight.ts:219

Per-layer verification results.


overallScore: number

Defined in: src/types/oversight.ts:221

Overall aggregated score (T-norm of layer scores, 0.0-1.0).


optional proofTraces?: ProofTraceDto[]

Defined in: src/types/oversight.ts:223

Failed proof traces with causal chains for violated constraints.


verdict: string

Defined in: src/types/oversight.ts:225

Overall verdict: “safe”, “unsafe”, “partial”, or “residuated”.


violations: FormalJudgeViolationDto[]

Defined in: src/types/oversight.ts:227

Specific constraint violations (if any).