diff --git a/docs/release-control/v6/internal/subsystems/api-contracts.md b/docs/release-control/v6/internal/subsystems/api-contracts.md
index da45fd1b7..91a851d52 100644
--- a/docs/release-control/v6/internal/subsystems/api-contracts.md
+++ b/docs/release-control/v6/internal/subsystems/api-contracts.md
@@ -467,7 +467,11 @@ the canonical monitored-system blocked payload.
    drawer handoffs may include live pending-approval metadata only as safe
    operator context: approval ID, status, risk, requested/expiry timestamps,
    target label, generated approval summary, and command count are allowed, while
-   approval command payloads stay inside governed approval/remediation surfaces
+   approval command payloads stay inside governed approval/remediation surfaces,
+   and finding-level first-turn prompt emphasis may be derived from the same
+   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
    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 607cc4534..d7490f773 100644
--- a/docs/release-control/v6/internal/subsystems/frontend-primitives.md
+++ b/docs/release-control/v6/internal/subsystems/frontend-primitives.md
@@ -821,8 +821,9 @@ frontend primitive boundary.
     model-only refresh, but the shared drawer stays a generic shell rather than
     a Patrol summary prompt builder. The Patrol helper may turn those same safe
     references into visible action labels, safety notes, and approval-aware
-    suggested prompts or first-turn prompt emphasis, but it must not expose raw
-    command or execution payloads.
+    suggested prompts or first-turn prompt emphasis for assessment and
+    finding-level handoffs, but it must not expose raw command or execution
+    payloads.
     The
     drawer may render a generic
     context-briefing band from `frontend-modern/src/stores/aiChat.ts`, but
diff --git a/docs/release-control/v6/internal/subsystems/patrol-intelligence.md b/docs/release-control/v6/internal/subsystems/patrol-intelligence.md
index e9885ba3a..066385e0b 100644
--- a/docs/release-control/v6/internal/subsystems/patrol-intelligence.md
+++ b/docs/release-control/v6/internal/subsystems/patrol-intelligence.md
@@ -161,7 +161,11 @@ Patrol-specific presentation helpers.
    not copy the approval command payload into Assistant drawer prose. The
    model-only runtime briefing must apply that same
    recovered approval reference when framing the operator decision and action
-   posture. Inline Patrol approval actions that open Assistant must follow the
+   posture. The initial visible prompt for a Patrol finding must lead with that
+   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. 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,
    attach the Patrol-owned visible drawer briefing for the pending approval, and
diff --git a/frontend-modern/src/components/AI/FindingsPanel.tsx b/frontend-modern/src/components/AI/FindingsPanel.tsx
index 6f4861132..112240714 100644
--- a/frontend-modern/src/components/AI/FindingsPanel.tsx
+++ b/frontend-modern/src/components/AI/FindingsPanel.tsx
@@ -492,6 +492,10 @@ export const FindingsPanel: Component<FindingsPanelProps> = (props) => {
       title,
       subject,
       description: finding.description,
+      investigationOutcome: finding.investigationOutcome,
+      remediationId: finding.remediationPlanId,
+      pendingApproval: pendingApprovalBriefing,
+      proposedFix,
       investigationRecord: finding.investigationRecord,
     });
     const briefing = buildPatrolAssistantFindingBriefing({
diff --git a/frontend-modern/src/components/AI/__tests__/FindingsPanel.test.ts b/frontend-modern/src/components/AI/__tests__/FindingsPanel.test.ts
index 13da25d62..c65dbcf6f 100644
--- a/frontend-modern/src/components/AI/__tests__/FindingsPanel.test.ts
+++ b/frontend-modern/src/components/AI/__tests__/FindingsPanel.test.ts
@@ -48,6 +48,8 @@ describe('FindingsPanel assistant handoff', () => {
     expect(findingsPanelSource).toContain('buildPatrolAssistantApprovalBriefingInput');
     expect(findingsPanelSource).toContain('buildPatrolAssistantProposedFixBriefingInput');
     expect(findingsPanelSource).toContain('briefing,');
+    expect(findingsPanelSource).toContain('investigationOutcome: finding.investigationOutcome');
+    expect(findingsPanelSource).toContain('remediationId: finding.remediationPlanId');
     expect(findingsPanelSource).toContain('investigationRecord: finding.investigationRecord');
     expect(findingsPanelSource).toContain('pendingApproval: pendingApprovalBriefing');
     expect(findingsPanelSource).toContain('proposedFix,');
diff --git a/frontend-modern/src/features/patrol/__tests__/patrolInvestigationContextModel.test.ts b/frontend-modern/src/features/patrol/__tests__/patrolInvestigationContextModel.test.ts
index 9b5e66cba..206545da5 100644
--- a/frontend-modern/src/features/patrol/__tests__/patrolInvestigationContextModel.test.ts
+++ b/frontend-modern/src/features/patrol/__tests__/patrolInvestigationContextModel.test.ts
@@ -455,6 +455,61 @@ describe('patrolInvestigationContextModel', () => {
     ).toContain('Use that record as the main context before suggesting next actions.');
   });
 
+  it('leads finding prompts with live governed approval context without command text', () => {
+    const prompt = buildPatrolAssistantFindingPrompt({
+      title: 'High CPU usage',
+      subject: 'web-server',
+      description: 'CPU stayed above 95%.',
+      pendingApproval: {
+        id: 'approval-1',
+        status: 'pending',
+        riskLevel: 'high',
+        targetName: 'web-server',
+        actionApprovalPolicy: 'operator',
+        actionPreflight: 'Service restart would be attempted after health checks.',
+        actionDryRunSummary: 'Dry run completed with one restart action.',
+      },
+      proposedFix: buildPatrolAssistantProposedFixBriefingInput({
+        description: 'Restart the workload service',
+        commands: ['systemctl restart workload.service'],
+        riskLevel: 'high',
+        targetHost: 'web-server',
+      }),
+    });
+
+    expect(prompt).toContain('Start by reviewing governed approval approval-1');
+    expect(prompt).toContain('approval status pending');
+    expect(prompt).toContain('high risk');
+    expect(prompt).toContain('approval policy attached');
+    expect(prompt).toContain('dry-run posture attached');
+    expect(prompt).toContain('safest next step');
+    expect(prompt).not.toContain('systemctl restart workload.service');
+  });
+
+  it('leads finding prompts with proposed-fix posture without command text', () => {
+    const prompt = buildPatrolAssistantFindingPrompt({
+      title: 'Nginx down',
+      subject: 'node-1',
+      description: 'The service stopped responding.',
+      investigationOutcome: 'fix_queued',
+      remediationId: 'remediation-1',
+      proposedFix: buildPatrolAssistantProposedFixBriefingInput({
+        description: 'Restart nginx',
+        commands: ['systemctl restart nginx'],
+        riskLevel: 'medium',
+        targetHost: 'node-1',
+        destructive: true,
+      }),
+    });
+
+    expect(prompt).toContain('Start by reviewing the governed proposed fix');
+    expect(prompt).toContain('medium risk');
+    expect(prompt).toContain('1 command recorded for approval context');
+    expect(prompt).toContain('destructive action');
+    expect(prompt).toContain('safest next step');
+    expect(prompt).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 659787f4e..d4db5a1ce 100644
--- a/frontend-modern/src/features/patrol/patrolInvestigationContextModel.ts
+++ b/frontend-modern/src/features/patrol/patrolInvestigationContextModel.ts
@@ -61,6 +61,10 @@ export interface PatrolAssistantFindingPromptInput {
   title: string;
   subject: string;
   description: string;
+  investigationOutcome?: string | null;
+  remediationId?: string | null;
+  pendingApproval?: PatrolAssistantApprovalBriefingInput | null;
+  proposedFix?: PatrolAssistantProposedFixBriefingInput | null;
   investigationRecord?: InvestigationRecord | null;
 }
 
@@ -343,18 +347,79 @@ export function buildPatrolAssistantFindingPrompt(
   const subject = normalizeText(input.subject) || 'the affected resource';
   const description = normalizeText(input.description);
   const hasRecord = Boolean(input.investigationRecord?.id);
+  const actionInstruction = buildPatrolAssistantFindingActionPromptInstruction(input);
 
   let prompt = `I'd like to discuss this Patrol finding: "${title}" on ${subject}.`;
   if (hasRecord) {
     prompt +=
       '\n\nPulse Patrol has a structured investigation record for this finding. Use that record as the main context before suggesting next actions.';
   }
+  if (actionInstruction) {
+    prompt += `\n\n${actionInstruction}`;
+  }
   if (description) {
     prompt += `\n\n${description}`;
   }
   return prompt;
 }
 
+function buildPatrolAssistantFindingActionPromptInstruction(
+  input: PatrolAssistantFindingPromptInput,
+): string | undefined {
+  const pendingApproval = normalizeApprovalBriefing(input.pendingApproval);
+  const record = buildPatrolInvestigationRecordPresentation(input.investigationRecord);
+  const proposedFix = record.proposedFix || normalizeProposedFixBriefing(input.proposedFix);
+  const remediationId = normalizeText(input.remediationId);
+  const outcome = normalizeText(input.investigationOutcome || input.investigationRecord?.outcome);
+  const normalizedOutcome = outcome.toLowerCase();
+
+  if (pendingApproval.id) {
+    const contextParts = [
+      pendingApproval.status ? `approval status ${pendingApproval.status}` : undefined,
+      pendingApproval.riskLevel ? `${pendingApproval.riskLevel} risk` : undefined,
+      pendingApproval.targetName
+        ? `target ${truncateContextText(pendingApproval.targetName, 120)}`
+        : undefined,
+      pendingApproval.actionApprovalPolicy ? 'approval policy attached' : undefined,
+      pendingApproval.actionPreflight || pendingApproval.actionDryRunSummary
+        ? 'dry-run posture attached'
+        : undefined,
+    ].filter(isNonEmptyString);
+
+    return `Start by reviewing governed approval ${pendingApproval.id}${
+      contextParts.length > 0 ? ` (${contextParts.join('; ')})` : ''
+    }. Use the attached Patrol context to explain prerequisites, risk, and the safest next step before any execution.`;
+  }
+
+  const hasGovernedActionPosture = Boolean(
+    proposedFix ||
+    remediationId ||
+    normalizeText(input.investigationRecord?.approval_id) ||
+    GOVERNED_ACTION_OUTCOMES.has(normalizedOutcome),
+  );
+  if (!hasGovernedActionPosture) {
+    return undefined;
+  }
+
+  const contextParts = [
+    proposedFix?.description
+      ? `proposed fix ${truncateContextText(proposedFix.description, 120)}`
+      : undefined,
+    proposedFix?.riskLabel ? `${proposedFix.riskLabel.toLowerCase()} risk` : undefined,
+    proposedFix?.targetHost
+      ? `target ${truncateContextText(proposedFix.targetHost, 120)}`
+      : undefined,
+    proposedFix?.commandSummary,
+    proposedFix?.destructive ? 'destructive action' : undefined,
+    remediationId ? `remediation ${truncateContextText(remediationId, 120)}` : undefined,
+    outcome ? `outcome ${formatIdentifierLabel(outcome)?.toLowerCase() || outcome}` : undefined,
+  ].filter(isNonEmptyString);
+
+  return `Start by reviewing the governed proposed fix or action posture${
+    contextParts.length > 0 ? ` (${contextParts.join('; ')})` : ''
+  }. Use the attached Patrol context to explain risk, prerequisites, and the safest next step without repeating command text.`;
+}
+
 export function buildPatrolAssessmentAssistantHandoff(
   input: PatrolAssessmentAssistantHandoffInput,
 ): PatrolAssessmentAssistantHandoff {