fix: detect effort changes from /effort command in ThinkingEffort widget

src/utils/jsonl-metadata.ts

@@ -8,6 +8,8 @@ export type TranscriptThinkingEffort = 'low' | 'medium' | 'high' | 'max';
 
 const MODEL_STDOUT_PREFIX = '<local-command-stdout>Set model to ';
 const MODEL_STDOUT_EFFORT_REGEX = /^<local-command-stdout>Set model to[\s\S]*? with (low|medium|high|max) effort<\/local-command-stdout>$/i;
+const EFFORT_STDOUT_PREFIX = '<local-command-stdout>Set effort level to ';
+const EFFORT_STDOUT_REGEX = /^<local-command-stdout>Set effort level to (low|medium|high|max)\b/i;
 
 interface TranscriptEntry { message?: { content?: string } }
 
@@ -44,6 +46,14 @@ export function getTranscriptThinkingEffort(transcriptPath: string | undefined):
             }
 
             const visibleContent = getVisibleText(entry.message.content).trim();
+
+            if (visibleContent.startsWith(EFFORT_STDOUT_PREFIX)) {
+                const effortMatch = EFFORT_STDOUT_REGEX.exec(visibleContent);
+                if (effortMatch) {
+                    return normalizeThinkingEffort(effortMatch[1]);
+                }
+            }
+
             if (!visibleContent.startsWith(MODEL_STDOUT_PREFIX)) {
                 continue;
             }

src/widgets/__tests__/ThinkingEffort.test.ts

@@ -27,6 +27,10 @@ const MODEL_WITH_HIGH_EFFORT = '<local-command-stdout>Set model to \u001b[1mopus
 const MODEL_WITH_LOW_EFFORT = '<local-command-stdout>Set model to \u001b[1msonnet (claude-sonnet-4-5)\u001b[22m with \u001b[1mlow\u001b[22m effort</local-command-stdout>';
 const MODEL_WITH_MAX_EFFORT = '<local-command-stdout>Set model to \u001b[1mopus (claude-opus-4-6)\u001b[22m with \u001b[1mmax\u001b[22m effort</local-command-stdout>';
 const MODEL_WITHOUT_EFFORT = '<local-command-stdout>Set model to \u001b[1msonnet (claude-sonnet-4-5)\u001b[22m</local-command-stdout>';
+const EFFORT_HIGH = '<local-command-stdout>Set effort level to \u001b[1mhigh\u001b[22m: Comprehensive implementation with extensive testing and documentation</local-command-stdout>';
+const EFFORT_LOW = '<local-command-stdout>Set effort level to \u001b[1mlow\u001b[22m: Quick, minimal-effort response</local-command-stdout>';
+const EFFORT_MEDIUM = '<local-command-stdout>Set effort level to \u001b[1mmedium\u001b[22m: Balanced response with good coverage</local-command-stdout>';
+const EFFORT_MAX = '<local-command-stdout>Set effort level to \u001b[1mmax\u001b[22m (this session only): Maximum capability with deepest reasoning (Opus 4.6 only)</local-command-stdout>';
 
 let tempDir: string;
 
@@ -155,6 +159,61 @@ describe('ThinkingEffortWidget', () => {
         });
     });
 
+    describe('/effort command source', () => {
+        it('reads effort from /effort transcript stdout', () => {
+            const result = render({ fileContent: makeTranscriptEntry(EFFORT_HIGH) });
+            expect(result).toBe('Thinking: high');
+        });
+
+        it('supports low effort from /effort command', () => {
+            const result = render({ fileContent: makeTranscriptEntry(EFFORT_LOW) });
+            expect(result).toBe('Thinking: low');
+        });
+
+        it('supports medium effort from /effort command', () => {
+            const result = render({ fileContent: makeTranscriptEntry(EFFORT_MEDIUM) });
+            expect(result).toBe('Thinking: medium');
+        });
+
+        it('supports max effort from /effort command', () => {
+            const result = render({ fileContent: makeTranscriptEntry(EFFORT_MAX) });
+            expect(result).toBe('Thinking: max');
+        });
+
+        it('returns raw effort from /effort command', () => {
+            const result = render({ fileContent: makeTranscriptEntry(EFFORT_HIGH), rawValue: true });
+            expect(result).toBe('high');
+        });
+
+        it('/effort overrides earlier /model when it is newer', () => {
+            const result = render({
+                fileContent: [
+                    makeTranscriptEntry(MODEL_WITH_LOW_EFFORT),
+                    makeTranscriptEntry(EFFORT_MAX)
+                ].join('\n')
+            });
+            expect(result).toBe('Thinking: max');
+        });
+
+        it('/model overrides earlier /effort when it is newer', () => {
+            const result = render({
+                fileContent: [
+                    makeTranscriptEntry(EFFORT_MAX),
+                    makeTranscriptEntry(MODEL_WITH_LOW_EFFORT)
+                ].join('\n')
+            });
+            expect(result).toBe('Thinking: low');
+        });
+
+        it('/effort overrides settings fallback', () => {
+            const result = render({
+                fileContent: makeTranscriptEntry(EFFORT_HIGH),
+                settingsValue: { effortLevel: 'low' }
+            });
+            expect(result).toBe('Thinking: high');
+        });
+    });
+
     describe('Claude settings fallback', () => {
         it('falls back to effortLevel when the latest /model output has no effort', () => {
             const result = render({