diff --git a/docs/release-control/v6/internal/subsystems/api-contracts.md b/docs/release-control/v6/internal/subsystems/api-contracts.md
index 7262f8cd4..c16ef506b 100644
--- a/docs/release-control/v6/internal/subsystems/api-contracts.md
+++ b/docs/release-control/v6/internal/subsystems/api-contracts.md
@@ -472,9 +472,10 @@ the canonical monitored-system blocked payload.
    bounded metadata only so Assistant starts by reviewing approval status, risk,
    dry-run posture, proposed-fix posture, and safest next step without receiving
    raw command or execution payloads; finding-level handoffs may also send one
-   bounded `handoff_actions` entry for that governed approval or proposed fix so
-   the Assistant runtime can refresh action posture from IDs and safe summaries
-   instead of relying on pasted chat text
+   bounded model-only `handoff_context`, one `handoff_resources` target
+   reference, and one `handoff_actions` entry for that governed approval or
+   proposed fix so the Assistant runtime can refresh finding and action posture
+   from IDs and safe summaries instead of relying on pasted chat text
    and the dedicated `frontend-modern/src/stores/aiIntelligenceSummaryModel.ts` owner, so recent-change counts and governed policy-posture fallbacks normalize once at the shared store boundary instead of as Patrol-hook-local payload repair
    and the shared `frontend-modern/src/components/Infrastructure/ResourceCorrelationSummary.tsx` card, so learned correlations and correlation context stay rendered through one governed frontend card instead of separate page-local list loops
    and the same shared correlation card's ordering and truncation rule, so callers pass raw correlations instead of encoding their own top-N sort behavior
diff --git a/docs/release-control/v6/internal/subsystems/frontend-primitives.md b/docs/release-control/v6/internal/subsystems/frontend-primitives.md
index f815e2026..38ad32b4d 100644
--- a/docs/release-control/v6/internal/subsystems/frontend-primitives.md
+++ b/docs/release-control/v6/internal/subsystems/frontend-primitives.md
@@ -823,6 +823,7 @@ frontend primitive boundary.
     references into visible action labels, safety notes, and approval-aware
     suggested prompts or first-turn prompt emphasis for assessment and
     finding-level handoffs; finding-level drawer opens may also pass one bounded
+    model-only finding context, one target resource reference, and one
     `handoff_actions` reference for a live approval or proposed fix. It must not
     expose raw command or execution payloads.
     The
diff --git a/docs/release-control/v6/internal/subsystems/patrol-intelligence.md b/docs/release-control/v6/internal/subsystems/patrol-intelligence.md
index 26ce1dc02..b6da291fe 100644
--- a/docs/release-control/v6/internal/subsystems/patrol-intelligence.md
+++ b/docs/release-control/v6/internal/subsystems/patrol-intelligence.md
@@ -165,10 +165,13 @@ Patrol-specific presentation helpers.
    governed approval or proposed-fix review instruction when safe metadata is
    attached, so the operator starts from approval status, risk, dry-run posture,
    and safest-next-step review instead of generic incident discussion. Finding
-   handoffs may also attach a bounded model-only action reference for that same
-   live approval or proposed fix, carrying approval ID/status/risk, action plan
-   identity/policy/expiry, dry-run posture, proposed-fix summary, and target
-   resource references without raw command payloads. Inline
+   handoffs must be assembled through the Patrol-owned handoff model so the
+   prompt, visible briefing, model-only finding context, resource reference,
+   bounded action reference, and request-local approval-required posture stay in
+   sync. The model-only context may include current finding status, recurrence,
+   investigation record facts, evidence, verification, approval posture,
+   dry-run posture, proposed-fix summary, and target resource references without
+   raw command payloads. Inline
    Patrol approval actions that open Assistant must follow the
    same rule: pass approval ID/status/risk/target plus safe summary/count
    metadata as review context, force the request-local approval-required mode,
diff --git a/frontend-modern/src/components/AI/FindingsPanel.tsx b/frontend-modern/src/components/AI/FindingsPanel.tsx
index cf789d54b..e9b76f131 100644
--- a/frontend-modern/src/components/AI/FindingsPanel.tsx
+++ b/frontend-modern/src/components/AI/FindingsPanel.tsx
@@ -18,10 +18,8 @@ import { aiIntelligenceStore, type UnifiedFinding } from '@/stores/aiIntelligenc
 import { notificationStore } from '@/stores/notifications';
 import { aiChatStore } from '@/stores/aiChat';
 import {
-  buildPatrolAssistantFindingBriefing,
-  buildPatrolAssistantFindingHandoffActions,
-  buildPatrolAssistantFindingPrompt,
   buildPatrolAssistantApprovalBriefingInput,
+  buildPatrolAssistantFindingHandoff,
   buildPatrolAssistantProposedFixBriefingInput,
   buildPatrolRemediationPlanAssistantBriefing,
   buildPatrolRemediationPlanAssistantPrompt,
@@ -489,61 +487,30 @@ export const FindingsPanel: Component<FindingsPanelProps> = (props) => {
     );
     const proposedFix =
       latestInvestigationProposedFix || buildLiveApprovalProposedFixBriefing(pendingApproval);
-    const prompt = buildPatrolAssistantFindingPrompt({
+    const handoff = buildPatrolAssistantFindingHandoff({
+      id: finding.id,
       title,
       subject,
       description: finding.description,
-      investigationOutcome: finding.investigationOutcome,
-      remediationId: finding.remediationPlanId,
-      pendingApproval: pendingApprovalBriefing,
-      proposedFix,
-      investigationRecord: finding.investigationRecord,
-    });
-    const briefing = buildPatrolAssistantFindingBriefing({
-      title,
-      subject,
       severity: finding.severity,
       findingStatus: finding.status,
+      investigationStatus: finding.investigationStatus,
       investigationOutcome: finding.investigationOutcome,
       loopState: finding.loopState,
       timesRaised: finding.timesRaised,
       regressionCount: finding.regressionCount,
       lastRegressionAt: finding.lastRegressionAt,
       remediationId: finding.remediationPlanId,
-      pendingApproval: pendingApprovalBriefing,
-      proposedFix,
-      investigationRecord: finding.investigationRecord,
-    });
-    const handoffActions = buildPatrolAssistantFindingHandoffActions({
-      id: finding.id,
-      title,
-      description: finding.description,
-      severity: finding.severity,
-      status: finding.status,
       resourceId: finding.resourceId,
       resourceName: finding.resourceName,
       resourceType: finding.resourceType,
-      investigationStatus: finding.investigationStatus,
-      investigationOutcome: finding.investigationOutcome,
-      loopState: finding.loopState,
+      detectedAt: finding.detectedAt,
+      lastSeenAt: finding.lastSeenAt,
       pendingApproval: pendingApprovalBriefing,
       proposedFix,
       investigationRecord: finding.investigationRecord,
     });
-    const requiresApprovalMode = patrolAssistantFindingHandoffRequiresApprovalMode({
-      investigationOutcome: finding.investigationOutcome,
-      remediationId: finding.remediationPlanId,
-      pendingApproval: pendingApprovalBriefing,
-      investigationRecord: finding.investigationRecord,
-    });
-    aiChatStore.openWithPrompt(prompt, {
-      targetType: finding.resourceType,
-      targetId: finding.resourceId,
-      findingId: finding.id,
-      briefing,
-      handoffActions: handoffActions.length > 0 ? handoffActions : undefined,
-      autonomousMode: requiresApprovalMode ? false : undefined,
-    });
+    aiChatStore.openWithPrompt(handoff.prompt, handoff.context);
   };
 
   const handleOpenPlanInAssistant = (finding: UnifiedFinding, plan: RemediationPlan, e: Event) => {
diff --git a/frontend-modern/src/components/AI/__tests__/FindingsPanel.test.ts b/frontend-modern/src/components/AI/__tests__/FindingsPanel.test.ts
index a09dd6e3a..a58e75aba 100644
--- a/frontend-modern/src/components/AI/__tests__/FindingsPanel.test.ts
+++ b/frontend-modern/src/components/AI/__tests__/FindingsPanel.test.ts
@@ -44,25 +44,23 @@ const patrolWorkspaceSource = readFileSync(
 
 describe('FindingsPanel assistant handoff', () => {
   it('routes Patrol investigation records into the Assistant briefing context', () => {
-    expect(findingsPanelSource).toContain('buildPatrolAssistantFindingBriefing');
-    expect(findingsPanelSource).toContain('buildPatrolAssistantFindingHandoffActions');
+    expect(findingsPanelSource).toContain('buildPatrolAssistantFindingHandoff');
     expect(findingsPanelSource).toContain('buildPatrolAssistantApprovalBriefingInput');
     expect(findingsPanelSource).toContain('buildPatrolAssistantProposedFixBriefingInput');
-    expect(findingsPanelSource).toContain('briefing,');
+    expect(findingsPanelSource).toContain(
+      'aiChatStore.openWithPrompt(handoff.prompt, handoff.context)',
+    );
     expect(findingsPanelSource).toContain('investigationOutcome: finding.investigationOutcome');
+    expect(findingsPanelSource).toContain('investigationStatus: finding.investigationStatus');
     expect(findingsPanelSource).toContain('remediationId: finding.remediationPlanId');
+    expect(findingsPanelSource).toContain('resourceId: finding.resourceId');
+    expect(findingsPanelSource).toContain('detectedAt: finding.detectedAt');
+    expect(findingsPanelSource).toContain('lastSeenAt: finding.lastSeenAt');
     expect(findingsPanelSource).toContain('investigationRecord: finding.investigationRecord');
     expect(findingsPanelSource).toContain('pendingApproval: pendingApprovalBriefing');
     expect(findingsPanelSource).toContain('proposedFix,');
-    expect(findingsPanelSource).toContain(
-      'handoffActions: handoffActions.length > 0 ? handoffActions : undefined',
-    );
     expect(findingsPanelSource).toContain('AIAPI.getInvestigation(finding.id)');
     expect(findingsPanelSource).toContain('await aiIntelligenceStore.loadPendingApprovals()');
-    expect(findingsPanelSource).toContain('patrolAssistantFindingHandoffRequiresApprovalMode');
-    expect(findingsPanelSource).toContain(
-      'autonomousMode: requiresApprovalMode ? false : undefined',
-    );
   });
 
   it('routes remediation plan handoffs through the command-free Patrol handoff model', () => {
diff --git a/frontend-modern/src/features/patrol/__tests__/patrolInvestigationContextModel.test.ts b/frontend-modern/src/features/patrol/__tests__/patrolInvestigationContextModel.test.ts
index eb66ec541..1b5cb2777 100644
--- a/frontend-modern/src/features/patrol/__tests__/patrolInvestigationContextModel.test.ts
+++ b/frontend-modern/src/features/patrol/__tests__/patrolInvestigationContextModel.test.ts
@@ -4,6 +4,7 @@ import type { RemediationPlan } from '@/api/ai';
 import {
   buildPatrolAssessmentAssistantHandoff,
   buildPatrolAssistantFindingBriefing,
+  buildPatrolAssistantFindingHandoff,
   buildPatrolAssistantFindingHandoffActions,
   buildPatrolAssistantFindingPrompt,
   buildPatrolAssistantProposedFixBriefingInput,
@@ -580,6 +581,76 @@ describe('patrolInvestigationContextModel', () => {
     ).toEqual([]);
   });
 
+  it('builds finding-level Assistant handoff context, resources, and actions together', () => {
+    const handoff = buildPatrolAssistantFindingHandoff({
+      id: 'finding-1',
+      title: 'Nginx down',
+      subject: 'node-1',
+      description: 'The service stopped responding.',
+      severity: 'warning',
+      findingStatus: 'active',
+      investigationStatus: 'completed',
+      investigationOutcome: 'fix_queued',
+      loopState: 'awaiting_approval',
+      timesRaised: 3,
+      regressionCount: 1,
+      lastRegressionAt: '2026-05-06T11:59:00Z',
+      remediationId: 'remediation-1',
+      resourceId: 'agent-1',
+      resourceName: 'node-1',
+      resourceType: 'agent',
+      detectedAt: '2026-05-06T11:50:00Z',
+      lastSeenAt: '2026-05-06T12:00:00Z',
+      pendingApproval: {
+        id: 'approval-1',
+        status: 'pending',
+        riskLevel: 'high',
+        requestedAt: '2026-05-06T12:00:00Z',
+        expiresAt: '2026-05-06T12:10:00Z',
+        targetName: 'node-1',
+        actionId: 'restart-nginx',
+        actionApprovalPolicy: 'operator',
+        actionPlanMessage: 'Restart nginx after validating load balancer drain.',
+        actionPreflight: 'Would restart nginx on node-1.',
+        actionDryRunSummary: 'One service restart would be attempted.',
+      },
+      proposedFix: buildPatrolAssistantProposedFixBriefingInput({
+        description: 'Restart nginx',
+        commands: ['systemctl restart nginx'],
+        riskLevel: 'high',
+        targetHost: 'node-1',
+      }),
+    });
+
+    expect(handoff.prompt).toContain('Start by reviewing governed approval approval-1');
+    expect(handoff.context).toMatchObject({
+      targetType: 'agent',
+      targetId: 'agent-1',
+      findingId: 'finding-1',
+      autonomousMode: false,
+      handoffResources: [{ id: 'agent-1', name: 'node-1', type: 'agent' }],
+      context: {
+        source: 'pulse-patrol-finding',
+        findingId: 'finding-1',
+        resourceId: 'agent-1',
+        resourceName: 'node-1',
+        resourceType: 'agent',
+        pendingApprovalId: 'approval-1',
+        actionReferenceCount: 1,
+      },
+    });
+    expect(handoff.context.handoffContext).toContain('[Patrol Finding Context]');
+    expect(handoff.context.handoffContext).toContain('Approval: approval-1');
+    expect(handoff.context.handoffContext).toContain(
+      'Dry-Run Posture: One service restart would be attempted.',
+    );
+    expect(handoff.context.handoffContext).toContain(
+      'Command Boundary: Command details stay in governed approval or remediation context',
+    );
+    expect(handoff.context.handoffActions).toHaveLength(1);
+    expect(JSON.stringify(handoff)).not.toContain('systemctl restart nginx');
+  });
+
   it('builds a drawer briefing for Assistant handoff without exposing raw commands', () => {
     const approvalRequestedAt = new Date(Date.now() - 60_000).toISOString();
     const approvalExpiresAt = new Date(Date.now() + 10 * 60_000).toISOString();
diff --git a/frontend-modern/src/features/patrol/patrolInvestigationContextModel.ts b/frontend-modern/src/features/patrol/patrolInvestigationContextModel.ts
index 1e618aa81..eeea82969 100644
--- a/frontend-modern/src/features/patrol/patrolInvestigationContextModel.ts
+++ b/frontend-modern/src/features/patrol/patrolInvestigationContextModel.ts
@@ -120,6 +120,30 @@ export interface PatrolAssistantFindingBriefingInput {
   investigationRecord?: InvestigationRecord | null;
 }
 
+export interface PatrolAssistantFindingHandoffInput {
+  id?: string | null;
+  title: string;
+  subject: string;
+  description?: string | null;
+  severity?: string | null;
+  findingStatus?: string | null;
+  investigationStatus?: string | null;
+  investigationOutcome?: string | null;
+  loopState?: string | null;
+  timesRaised?: number | null;
+  regressionCount?: number | null;
+  lastRegressionAt?: string | null;
+  remediationId?: string | null;
+  resourceId?: string | null;
+  resourceName?: string | null;
+  resourceType?: string | null;
+  detectedAt?: string | null;
+  lastSeenAt?: string | null;
+  pendingApproval?: PatrolAssistantApprovalBriefingInput | null;
+  proposedFix?: PatrolAssistantProposedFixBriefingInput | null;
+  investigationRecord?: InvestigationRecord | null;
+}
+
 export interface PatrolAssistantFindingModeInput {
   investigationOutcome?: string | null;
   remediationId?: string | null;
@@ -206,6 +230,11 @@ export interface PatrolAssessmentAssistantHandoff {
   context: Omit<AIChatContext, 'initialPrompt'>;
 }
 
+export interface PatrolAssistantFindingHandoff {
+  prompt: string;
+  context: Omit<AIChatContext, 'initialPrompt'>;
+}
+
 const MAX_ASSESSMENT_FINDINGS = 5;
 const MAX_ASSESSMENT_RECENT_CHANGES = 3;
 const MAX_ASSESSMENT_CORRELATIONS = 3;
@@ -454,6 +483,68 @@ export function buildPatrolAssessmentAssistantHandoff(
   };
 }
 
+export function buildPatrolAssistantFindingHandoff(
+  input: PatrolAssistantFindingHandoffInput,
+): PatrolAssistantFindingHandoff {
+  const findingId = normalizeText(input.id) || normalizeText(input.investigationRecord?.finding_id);
+  const resource = buildPatrolFindingHandoffResource(input);
+  const handoffResources = resource ? [resource] : [];
+  const handoffActions = buildPatrolAssistantFindingHandoffActions(input);
+  const requiresApprovalMode = patrolAssistantFindingHandoffRequiresApprovalMode({
+    investigationOutcome: input.investigationOutcome,
+    remediationId: input.remediationId,
+    pendingApproval: input.pendingApproval,
+    investigationRecord: input.investigationRecord,
+  });
+
+  return {
+    prompt: buildPatrolAssistantFindingPrompt({
+      title: input.title,
+      subject: input.subject,
+      description: normalizeText(input.description),
+      investigationOutcome: input.investigationOutcome,
+      remediationId: input.remediationId,
+      pendingApproval: input.pendingApproval,
+      proposedFix: input.proposedFix,
+      investigationRecord: input.investigationRecord,
+    }),
+    context: {
+      targetType: resource?.type,
+      targetId: resource?.id,
+      findingId: findingId || undefined,
+      autonomousMode: requiresApprovalMode ? false : undefined,
+      handoffContext: buildPatrolAssistantFindingModelContext(input),
+      handoffResources: handoffResources.length > 0 ? handoffResources : undefined,
+      handoffActions: handoffActions.length > 0 ? handoffActions : undefined,
+      briefing: buildPatrolAssistantFindingBriefing({
+        title: input.title,
+        subject: input.subject,
+        severity: input.severity,
+        findingStatus: input.findingStatus,
+        investigationOutcome: input.investigationOutcome,
+        loopState: input.loopState,
+        timesRaised: input.timesRaised,
+        regressionCount: input.regressionCount,
+        lastRegressionAt: input.lastRegressionAt,
+        remediationId: input.remediationId,
+        pendingApproval: input.pendingApproval,
+        proposedFix: input.proposedFix,
+        investigationRecord: input.investigationRecord,
+      }),
+      context: {
+        source: 'pulse-patrol-finding',
+        findingId: findingId || undefined,
+        investigationRecordId: normalizeText(input.investigationRecord?.id) || undefined,
+        resourceId: resource?.id,
+        resourceName: resource?.name,
+        resourceType: resource?.type,
+        pendingApprovalId: normalizeApprovalBriefing(input.pendingApproval).id || undefined,
+        actionReferenceCount: handoffActions.length,
+      },
+    },
+  };
+}
+
 export function buildPatrolAssistantFindingHandoffActions(
   finding: PatrolAssessmentAssistantFindingInput,
 ): AIChatHandoffAction[] {
@@ -762,6 +853,25 @@ function buildPatrolAssessmentHandoffResources(
   return Array.from(resources.values());
 }
 
+function buildPatrolFindingHandoffResource(
+  input: PatrolAssistantFindingHandoffInput,
+): AIChatHandoffResource | undefined {
+  const subject = input.investigationRecord?.subject;
+  const id = normalizeText(input.resourceId) || normalizeText(subject?.resource_id);
+  if (!id) return undefined;
+
+  return {
+    id,
+    name:
+      normalizeText(input.resourceName) ||
+      normalizeText(subject?.resource_name) ||
+      normalizeText(input.subject) ||
+      undefined,
+    type: normalizeText(input.resourceType) || normalizeText(subject?.resource_type) || undefined,
+    node: normalizeText(subject?.node) || undefined,
+  };
+}
+
 function addAssessmentHandoffResource(
   resources: Map<string, AIChatHandoffResource>,
   resource: AIChatHandoffResource,
@@ -1243,6 +1353,99 @@ function formatAssessmentFindingContextLine(
   return `Finding ${index}: ${parts.join('; ')}`;
 }
 
+function buildPatrolAssistantFindingModelContext(
+  input: PatrolAssistantFindingHandoffInput,
+): string {
+  const title = normalizeText(input.title) || 'Patrol finding';
+  const subject = normalizeText(input.subject) || 'affected resource';
+  const record = buildPatrolInvestigationRecordPresentation(input.investigationRecord);
+  const pendingApproval = normalizeApprovalBriefing(input.pendingApproval);
+  const proposedFix = record.proposedFix || normalizeProposedFixBriefing(input.proposedFix);
+  const resource = buildPatrolFindingHandoffResource(input);
+  const findingId = normalizeText(input.id) || normalizeText(input.investigationRecord?.finding_id);
+  const statusParts = [
+    formatIdentifierLabel(input.severity),
+    formatIdentifierLabel(input.findingStatus),
+    formatIdentifierLabel(input.investigationStatus),
+    formatIdentifierLabel(input.investigationOutcome || input.investigationRecord?.outcome),
+    formatIdentifierLabel(input.loopState),
+  ].filter(isNonEmptyString);
+  const raisedParts = [
+    normalizeNonNegativeCount(input.timesRaised) > 1
+      ? `raised ${normalizeNonNegativeCount(input.timesRaised)} times`
+      : undefined,
+    normalizeNonNegativeCount(input.regressionCount) > 0
+      ? `regressed ${normalizeNonNegativeCount(input.regressionCount)} time${
+          normalizeNonNegativeCount(input.regressionCount) === 1 ? '' : 's'
+        }`
+      : undefined,
+    normalizeText(input.lastRegressionAt)
+      ? `last regression ${normalizeText(input.lastRegressionAt)}`
+      : undefined,
+  ].filter(isNonEmptyString);
+  const attentionReason = buildPatrolAssistantAttentionReason(input, record);
+  const operatorDecision = buildPatrolAssistantOperatorDecision(input);
+  const proposedFixFacts = proposedFix
+    ? formatBriefingStringList(
+        [
+          proposedFix.description,
+          proposedFix.targetHost ? `target ${proposedFix.targetHost}` : undefined,
+          proposedFix.riskLabel ? `${proposedFix.riskLabel.toLowerCase()} risk` : undefined,
+          proposedFix.commandSummary,
+          proposedFix.destructive ? 'destructive proposed fix' : undefined,
+          proposedFix.rationale ? `rationale ${proposedFix.rationale}` : undefined,
+        ],
+        6,
+        'proposed-fix facts',
+      )
+    : undefined;
+
+  return [
+    '[Patrol Finding Context]',
+    'Source: Pulse Patrol finding handoff',
+    formatContextLine('Finding', title),
+    formatContextLine('Finding ID', findingId),
+    formatContextLine('Subject', subject),
+    formatContextLine('Resource', resource ? formatAssessmentResourceLabel(resource) : undefined),
+    formatContextLine('Status', statusParts.join(' · ')),
+    formatContextLine('Detected At', input.detectedAt),
+    formatContextLine('Last Seen At', input.lastSeenAt),
+    formatContextLine('Recurrence', raisedParts.join('; ')),
+    formatContextLine('Description', input.description),
+    formatContextLine('Attention', attentionReason),
+    formatContextLine('Investigation Record', input.investigationRecord?.id),
+    formatContextLine('Investigation Status', record.statusLabel),
+    formatContextLine('Investigation Outcome', record.outcomeLabel),
+    formatContextLine('Investigation Confidence', record.confidenceLabel),
+    formatContextLine('Conclusion', record.conclusion),
+    formatContextLine('Recommended Action', record.recommendedAction),
+    ...record.evidenceSummaries.map((summary, index) =>
+      formatContextLine(`Evidence ${index + 1}`, summary),
+    ),
+    ...record.verificationSummaries.map((summary, index) =>
+      formatContextLine(`Verification ${index + 1}`, summary),
+    ),
+    formatContextLine('Tools Used', record.toolsUsed.join(', ')),
+    formatContextLine('Approval', pendingApproval.id),
+    formatContextLine('Approval Status', pendingApproval.status),
+    formatContextLine('Approval Risk', pendingApproval.riskLevel),
+    formatContextLine('Approval Target', pendingApproval.targetName),
+    formatContextLine('Approval Requested At', pendingApproval.requestedAt),
+    formatContextLine('Approval Expires At', pendingApproval.expiresAt),
+    formatContextLine('Approval Policy', pendingApproval.actionApprovalPolicy),
+    formatContextLine('Approval Plan Expires At', pendingApproval.actionPlanExpiresAt),
+    formatContextLine('Action Plan Summary', pendingApproval.actionPlanMessage),
+    formatContextLine('Action Preflight', pendingApproval.actionPreflight),
+    formatContextLine('Dry-Run Posture', pendingApproval.actionDryRunSummary),
+    formatContextLine('Proposed Fix', proposedFixFacts),
+    formatContextLine('Operator Decision', operatorDecision),
+    'Command Boundary: Command details stay in governed approval or remediation context; this model-only handoff may include command counts but not raw command text.',
+    'Operator Boundary: This Patrol finding handoff is model-only context for explanation and review. Diagnostics, remediation, and command execution require explicit governed approval.',
+  ]
+    .filter(isNonEmptyString)
+    .join('\n');
+}
+
 function formatAssessmentPendingApprovalContextParts(
   approval: Required<PatrolAssistantApprovalBriefingInput>,
 ): string[] {