LegalLegal AI Ops Trust Overview

Use this workspace to inspect trust movement across legal reasoning, structured review, and policy-bound outputs. Graph rows reflect fleet trust-state data; the badge is a reading lens for this vertical.

TrustGraph

Live Trust State

Reflex Score: 0.82 · Recovery Active: false

Loading trust indicator…

← Classic Trust Graph·Surface Alignment Ω.2 · Runtime (read-only)·Replay Studio (same mission scope)

Mission verification

Executive verification surface — mission intent, evidence coverage, and trust posture at a glance. Advanced TrustGraph runtime tooling stays collapsed until you need it.

Mission: mission_test_001

Mission identifier

Verifying…

Current phase

—

Risk posture

Verifying…

Proof coverage

Pending

Verification

Verifying…

Material risk

Verifying…

Verification status

Verifying…

Return to Workspace View mission evidence Open advanced TrustGraph

Focused lineage

First Contact intent
Mission bound
Trust baseline
Safeguards confirmed
Verification aligned
Current evidence state

Return to Workspace View Evidence Open Builder Review

Mission evidence

Loading evidence summary…

Coverage review summary

Builder coverage review

Pending human review

8 gaps identified

Open Builder review

Builder readiness

Awaiting human review

8 coverage gaps | 8 helper proposals | 14 files | 8 required tests

Review in Mission Builder

Proposed implementation package

8 helper proposals | 14 files | 8 tests

Trust Lock constraints preserved (9) | Execution authority: none

Open task pack

Advanced Topology

Scope: mission_test_001

Advanced verification controls

Replay boundary

Active doctrine for this TrustGraph runtime surface (Ω.2 — canonical ontology registry).

Replay-only
Visualization rebuilds from fetched trust-state window; no live mutation.
Append-only
Fleet history grows forward; this surface does not rewrite prior records.
Non-enforcing
No orchestration or policy execution from this canvas.
Execution-authority separated
Trust signals are inspectable; runtime control remains outside this view.
Does not mutate runtime execution
Mission-builder projection and graph inspection do not alter live execution or deploy paths.

Verifier exchange overlay

Optional pasted bundle commits reference digests; topology is not fetched outside the bundle envelope.

Open Replay Studio export →Guided Replay Studio demo →

Org context · org_demo_legal· Mission · mission_test_001

Paste bundle JSON exported from Replay Studio. JSON is parsed only — nothing is executed from the clipboard.

· Tenant-scoped previews only — no ontology or doctrine mutation from this UI.
· Paste bundle JSON locally; no verifier network round-trip from this strip.

Advanced runtime diagnostics

Runtime diagnostics

Continuity: Continuity signals observed

Verification: Verifier lane activity observed

Lineage: Mission binding present

Survivability: Survivability cluster mapped

Topology intelligence

TrustGraph · runtime cognition

djb2:1aaa0f3a

Runtime cognition — posture summaries derived from projection only; read-only; no execution authority.

Continuity posture

Coverage coordination facet (advisory label)

Continuity posture — mesh facets present (advisory) · Coverage coordination facet (advisory label)

Verification posture

Verification posture — no verifier lane in projectionVerifier lane absent — integrity observation idle in this projection.Integrity posture observation — awaiting verifier lane in projection.

Verification posture — no verifier lane in projection · Verifier lane absent — integrity observation idle in this projection. · Integrity posture observation — awaiting verifier lane in projection.

Lineage posture

Lineage posture — slice stability echo (advisory)Mission lineage bundle identifier buckets (counts only).

Lineage posture — slice stability echo (advisory) · Mission lineage bundle identifier buckets (counts only). · Bundle counts · 1 task pack identifiers in slice · 0 coverage decision identifiers in slice · 0 receipt identifiers in slice · 0 review record identifiers in slice

Survivability posture

Survivability posture — no readiness headline or audit echo in projectionRegistry verification is verifier-first; ribbon metrics are observational — no merge, deploy, or orchestration actions on this surface.

Survivability posture — no readiness headline or audit echo in projection · Registry verification is verifier-first; ribbon metrics are observational — no merge, deploy, or orchestration actions on this surface.

Coherence memory (advisory)

TrustGraph · runtime coherence memory

Coherence summary — longitudinal facets compose cognition timeline and projection digest only.

Cognition drift · insufficient temporal span for longitudinal read

Drift · Coherence band · unknown / single-point (advisory)Continuity stability · Coherence band · unknown / single-point (advisory)Verifier consistency · Coherence band · unknown / single-point (advisory)Survivability evolution · Coherence band · unknown / single-point (advisory)

Raw diagnostics (engineers)

Continuity: runtimeTopologyIntelligenceContinuityWorstBandStreakShort · runtimeTopologyIntelligenceContinuityRiskSignatureInsufficientSpan · runtimeTopologyIntelligenceContinuityRunTopologySparse
Verification: runtimeTopologyIntelligenceVerifierAnomalyTopologySparse
Survivability: runtimeTrustTopologySurvivabilityClusterSparse
Derived projection — read-only; supplements the selection mesh; not an execution or persistence source.
Mission lineage slice · org_demo_legal · mission_test_001 · 1 task pack identifiers in slice · 0 coverage decision identifiers in slice · 0 receipt identifiers in slice · 0 review record identifiers in slice
Slice identifiers are advisory lineage pointers — execution authority stays separated.

TrustGraph canvas

Fleet: Degraded · Reflex 0.82 · Recovery inactive

Summary metrics appear when trust edges exist. Partner demo mode uses defensibility-oriented labels; underlying counts are unchanged.

Local trust (nodes & edges)

healthy

degraded

critical

Recovery is a global fleet overlay (nav / indicator API). It tints emphasis on the canvas but is not assigned to individual nodes from local scores.

Edge labels with a sky-tinted pill are cache / velocity metadata (recommended refresh + domain speed). They do not change propagation enforcement — stroke color still follows edgeAction when present.

Nodes use the worst (minimum) reflex seen on any attached trust observation. Merged links keep the minimum reflex across combined rows.

No trust states yet — graph appears when edges exist.

Recent graph edges

Counsel lane	Validation step	Policy surface	Result type	Score
No trust states yet

Edge inspector

System-wide trust

Degraded

Global reflex: 0.82

Recovery active: false

Loading trust indicator…

Trust path (selected)

Select a row to compare edge-level reflex with fleet trust. Underlying agent, tool, and workflow names are shown below each lens label.

Partner demo · recovery arc

Legal — citation to review-safe handoff

Narrative lens only—graph rows and scores above are unchanged system data.

Citation lineage breaks on a high-stakes clause.
Defensibility risk rises on the contract-review path.
Recovery adds verification before clause extraction.
Output stabilizes as review-safe with traceable lineage.

Recommended action

Strengthen defensibility checks before clause extraction.

Operational detail: Investigate citation and policy drift; tighten monitoring on high-risk matters and contract handoffs.