refactor(eval): inject logger dependency into evaluators and runners

lefthook.yml

@@ -14,6 +14,9 @@ prepare-commit-msg:
   skip:
     - merge
     - rebase
-  jobs:
-    - run: ./dist/cli.js --message-only > {1}
+  commands:
+    commitment:
+      # Only run for regular commits (not merge, squash, or when message specified)
+      # {1} is the commit message file, {2} is the commit source
+      run: '[ -z "{2}" ] && ./dist/cli.js --message-only > {1} || true'
       interactive: true

src/eval/evaluators/__tests__/chatgpt-agent.test.ts

@@ -10,6 +10,7 @@
 
 import { beforeEach, describe, expect, it, mock } from 'bun:test';
 import { z } from 'zod';
+import { SilentLogger } from '../../../utils/logger.js';
 import { ChatGPTAgent } from '../chatgpt-agent.js';
 
 // Mock the OpenAI Agents SDK
@@ -30,7 +31,7 @@ describe('ChatGPTAgent', () => {
   describe('evaluate()', () => {
     it('should use gpt-5 model', async () => {
       const schema = z.object({ result: z.string() });
-      const agent = new ChatGPTAgent();
+      const agent = new ChatGPTAgent(new SilentLogger());
 
       mockAgent.mockReturnValue({});
       mockRun.mockResolvedValue({
@@ -49,7 +50,7 @@ describe('ChatGPTAgent', () => {
 
     it('should use outputType pattern with Zod schema', async () => {
       const schema = z.object({ score: z.number() });
-      const agent = new ChatGPTAgent();
+      const agent = new ChatGPTAgent(new SilentLogger());
 
       mockAgent.mockReturnValue({});
       mockRun.mockResolvedValue({
@@ -68,7 +69,7 @@ describe('ChatGPTAgent', () => {
 
     it('should access data via result.finalOutput', async () => {
       const schema = z.object({ data: z.string() });
-      const agent = new ChatGPTAgent();
+      const agent = new ChatGPTAgent(new SilentLogger());
 
       mockAgent.mockReturnValue({});
       mockRun.mockResolvedValue({
@@ -82,7 +83,7 @@ describe('ChatGPTAgent', () => {
 
     it('should pass instructions to Agent', async () => {
       const schema = z.object({ value: z.number() });
-      const agent = new ChatGPTAgent();
+      const agent = new ChatGPTAgent(new SilentLogger());
       const instructions = 'Evaluate on scale 0-10';
 
       mockAgent.mockReturnValue({});
@@ -101,7 +102,7 @@ describe('ChatGPTAgent', () => {
 
     it('should include agent name in configuration', async () => {
       const schema = z.object({ result: z.boolean() });
-      const agent = new ChatGPTAgent();
+      const agent = new ChatGPTAgent(new SilentLogger());
 
       mockAgent.mockReturnValue({});
       mockRun.mockResolvedValue({
@@ -119,7 +120,7 @@ describe('ChatGPTAgent', () => {
 
     it('should throw EvaluationError on API failure', async () => {
       const schema = z.object({ result: z.string() });
-      const agent = new ChatGPTAgent();
+      const agent = new ChatGPTAgent(new SilentLogger());
       const apiError = new Error('API timeout');
 
       mockAgent.mockReturnValue({});
@@ -132,7 +133,7 @@ describe('ChatGPTAgent', () => {
 
     it('should handle missing finalOutput', async () => {
       const schema = z.object({ result: z.string() });
-      const agent = new ChatGPTAgent();
+      const agent = new ChatGPTAgent(new SilentLogger());
 
       mockAgent.mockReturnValue({});
       mockRun.mockResolvedValue({
@@ -146,7 +147,7 @@ describe('ChatGPTAgent', () => {
       const schema = z.object({
         score: z.number().min(0).max(10),
       });
-      const agent = new ChatGPTAgent();
+      const agent = new ChatGPTAgent(new SilentLogger());
 
       mockAgent.mockReturnValue({});
       // Simulate OpenAI returning invalid data that fails schema validation

src/eval/evaluators/__tests__/meta-evaluator.test.ts

@@ -9,6 +9,7 @@
  */
 
 import { beforeEach, describe, expect, it, mock } from 'bun:test';
+import { SilentLogger } from '../../../utils/logger.js';
 import type { AttemptOutcome } from '../../core/types.js';
 import { MetaEvaluator } from '../meta-evaluator.js';
 
@@ -18,6 +19,8 @@ const mockEvaluate = mock();
 mock.module('../chatgpt-agent.js', () => ({
   // biome-ignore lint/style/useNamingConvention: Mock needs to match exported class name
   ChatGPTAgent: class MockChatGPTAgent {
+    // biome-ignore lint/complexity/noUselessConstructor: Mock needs constructor for logger parameter
+    constructor(_logger: any) {} // Accept logger parameter
     evaluate = mockEvaluate;
   },
 }));
@@ -28,7 +31,7 @@ describe('MetaEvaluator', () => {
   });
   describe('evaluate() - 3/3 success', () => {
     it('should evaluate all 3 successful attempts', async () => {
-      const evaluator = new MetaEvaluator();
+      const evaluator = new MetaEvaluator(new SilentLogger());
       const attempts: AttemptOutcome[] = [
         {
           attemptNumber: 1,
@@ -73,7 +76,7 @@ describe('MetaEvaluator', () => {
     });
 
     it('should have high consistency for similar scores', async () => {
-      const evaluator = new MetaEvaluator();
+      const evaluator = new MetaEvaluator(new SilentLogger());
       const attempts: AttemptOutcome[] = [
         {
           attemptNumber: 1,
@@ -118,7 +121,7 @@ describe('MetaEvaluator', () => {
 
   describe('evaluate() - 2/3 success', () => {
     it('should penalize failures in finalScore', async () => {
-      const evaluator = new MetaEvaluator();
+      const evaluator = new MetaEvaluator(new SilentLogger());
       const attempts: AttemptOutcome[] = [
         {
           attemptNumber: 1,
@@ -164,7 +167,7 @@ describe('MetaEvaluator', () => {
     });
 
     it('should identify best attempt among successes', async () => {
-      const evaluator = new MetaEvaluator();
+      const evaluator = new MetaEvaluator(new SilentLogger());
       const attempts: AttemptOutcome[] = [
         {
           attemptNumber: 1,
@@ -208,7 +211,7 @@ describe('MetaEvaluator', () => {
 
   describe('evaluate() - 1/3 success', () => {
     it('should heavily penalize 2 failures', async () => {
-      const evaluator = new MetaEvaluator();
+      const evaluator = new MetaEvaluator(new SilentLogger());
       const attempts: AttemptOutcome[] = [
         {
           attemptNumber: 1,
@@ -251,7 +254,7 @@ describe('MetaEvaluator', () => {
     });
 
     it('should set consistency to 0 with only 1 success', async () => {
-      const evaluator = new MetaEvaluator();
+      const evaluator = new MetaEvaluator(new SilentLogger());
       const attempts: AttemptOutcome[] = [
         {
           attemptNumber: 1,
@@ -294,7 +297,7 @@ describe('MetaEvaluator', () => {
 
   describe('evaluate() - 0/3 success', () => {
     it('should provide reasoning even with all failures', async () => {
-      const evaluator = new MetaEvaluator();
+      const evaluator = new MetaEvaluator(new SilentLogger());
       const attempts: AttemptOutcome[] = [
         {
           attemptNumber: 1,
@@ -338,7 +341,7 @@ describe('MetaEvaluator', () => {
     });
 
     it('should set bestAttempt to undefined', async () => {
-      const evaluator = new MetaEvaluator();
+      const evaluator = new MetaEvaluator(new SilentLogger());
       const attempts: AttemptOutcome[] = [
         {
           attemptNumber: 1,
@@ -380,7 +383,7 @@ describe('MetaEvaluator', () => {
 
   describe('validate inputs', () => {
     it('should throw on invalid attempt count', async () => {
-      const evaluator = new MetaEvaluator();
+      const evaluator = new MetaEvaluator(new SilentLogger());
       const attempts: AttemptOutcome[] = [
         {
           attemptNumber: 1,
@@ -405,7 +408,7 @@ describe('MetaEvaluator', () => {
     });
 
     it('should handle ChatGPT API errors', async () => {
-      const evaluator = new MetaEvaluator();
+      const evaluator = new MetaEvaluator(new SilentLogger());
       const attempts: AttemptOutcome[] = [
         {
           attemptNumber: 1,
@@ -441,7 +444,7 @@ describe('MetaEvaluator', () => {
 
   describe('build comprehensive prompt', () => {
     it('should include all attempts in prompt', async () => {
-      const evaluator = new MetaEvaluator();
+      const evaluator = new MetaEvaluator(new SilentLogger());
       const attempts: AttemptOutcome[] = [
         {
           attemptNumber: 1,

src/eval/evaluators/__tests__/single-attempt.test.ts

@@ -9,6 +9,7 @@
  */
 
 import { beforeEach, describe, expect, it, mock } from 'bun:test';
+import { SilentLogger } from '../../../utils/logger.js';
 import { SingleAttemptEvaluator } from '../single-attempt.js';
 
 // Mock ChatGPTAgent
@@ -17,6 +18,8 @@ const mockEvaluate = mock();
 mock.module('../chatgpt-agent.js', () => ({
   // biome-ignore lint/style/useNamingConvention: Mock needs to match exported class name
   ChatGPTAgent: class MockChatGPTAgent {
+    // biome-ignore lint/complexity/noUselessConstructor: Mock needs constructor for logger parameter
+    constructor(_logger: any) {} // Accept logger parameter
     evaluate = mockEvaluate;
   },
 }));
@@ -27,7 +30,7 @@ describe('SingleAttemptEvaluator', () => {
   });
   describe('evaluate()', () => {
     it('should evaluate commit message with 4 metrics', async () => {
-      const evaluator = new SingleAttemptEvaluator();
+      const evaluator = new SingleAttemptEvaluator(new SilentLogger());
       const mockMetrics = {
         clarity: 9,
         conventionalFormat: 10,
@@ -47,7 +50,7 @@ describe('SingleAttemptEvaluator', () => {
     });
 
     it('should calculate overall score as average of metrics', async () => {
-      const evaluator = new SingleAttemptEvaluator();
+      const evaluator = new SingleAttemptEvaluator(new SilentLogger());
       const mockMetrics = {
         clarity: 8,
         conventionalFormat: 9,
@@ -64,7 +67,7 @@ describe('SingleAttemptEvaluator', () => {
     });
 
     it('should pass commit message to ChatGPT', async () => {
-      const evaluator = new SingleAttemptEvaluator();
+      const evaluator = new SingleAttemptEvaluator(new SilentLogger());
       const commitMessage = 'feat(api): add user endpoint';
 
       mockEvaluate.mockResolvedValue({
@@ -83,7 +86,7 @@ describe('SingleAttemptEvaluator', () => {
     });
 
     it('should include diff in evaluation context', async () => {
-      const evaluator = new SingleAttemptEvaluator();
+      const evaluator = new SingleAttemptEvaluator(new SilentLogger());
       const diff = 'diff --git a/src/api.ts b/src/api.ts\n+new code';
 
       mockEvaluate.mockResolvedValue({
@@ -102,7 +105,7 @@ describe('SingleAttemptEvaluator', () => {
     });
 
     it('should include fixture name in context', async () => {
-      const evaluator = new SingleAttemptEvaluator();
+      const evaluator = new SingleAttemptEvaluator(new SilentLogger());
       const fixtureName = 'complex-refactoring';
 
       mockEvaluate.mockResolvedValue({
@@ -121,7 +124,7 @@ describe('SingleAttemptEvaluator', () => {
     });
 
     it('should validate metrics are in 0-10 range', async () => {
-      const evaluator = new SingleAttemptEvaluator();
+      const evaluator = new SingleAttemptEvaluator(new SilentLogger());
 
       // Simulate ChatGPT returning invalid metrics that fail schema validation
       mockEvaluate.mockRejectedValue(
@@ -132,15 +135,15 @@ describe('SingleAttemptEvaluator', () => {
     });
 
     it('should handle ChatGPT evaluation errors', async () => {
-      const evaluator = new SingleAttemptEvaluator();
+      const evaluator = new SingleAttemptEvaluator(new SilentLogger());
 
       mockEvaluate.mockRejectedValue(new Error('API timeout'));
 
       await expect(evaluator.evaluate('message', 'diff', 'fixture')).rejects.toThrow();
     });
 
     it('should handle edge case: all metrics are 10', async () => {
-      const evaluator = new SingleAttemptEvaluator();
+      const evaluator = new SingleAttemptEvaluator(new SilentLogger());
 
       mockEvaluate.mockResolvedValue({
         clarity: 10,
@@ -155,7 +158,7 @@ describe('SingleAttemptEvaluator', () => {
     });
 
     it('should handle edge case: all metrics are 0', async () => {
-      const evaluator = new SingleAttemptEvaluator();
+      const evaluator = new SingleAttemptEvaluator(new SilentLogger());
 
       mockEvaluate.mockResolvedValue({
         clarity: 0,
@@ -170,7 +173,7 @@ describe('SingleAttemptEvaluator', () => {
     });
 
     it('should round overall score to 1 decimal place', async () => {
-      const evaluator = new SingleAttemptEvaluator();
+      const evaluator = new SingleAttemptEvaluator(new SilentLogger());
 
       mockEvaluate.mockResolvedValue({
         clarity: 7,

src/eval/evaluators/chatgpt-agent.ts

@@ -32,6 +32,7 @@
 
 import type { AgentOutputType } from '@openai/agents';
 import { Agent, run } from '@openai/agents';
+import type { Logger } from '../../utils/logger.js';
 import { EvaluationError } from '../core/errors.js';
 
 /**
@@ -41,6 +42,15 @@ import { EvaluationError } from '../core/errors.js';
  * and returns typed results via outputType pattern.
  */
 export class ChatGPTAgent {
+  /**
+   * Create a new ChatGPT agent
+   *
+   * @param _logger - Logger for progress messages (reserved for future use)
+   */
+  constructor(_logger: Logger) {
+    // Logger reserved for future use
+    void _logger;
+  }
   /**
    * Evaluate using ChatGPT with structured output
    *

src/eval/evaluators/meta-evaluator.ts

@@ -28,6 +28,7 @@
  * ```
  */
 
+import type { Logger } from '../../utils/logger.js';
 import { EvaluationError } from '../core/errors.js';
 import { metaEvaluationOutputSchema } from '../core/schemas.js';
 import type { AttemptOutcome, EvalResult } from '../core/types.js';
@@ -44,9 +45,11 @@ export class MetaEvaluator {
 
   /**
    * Create a new meta-evaluator
+   *
+   * @param logger - Logger for progress messages (reserved for future use)
    */
-  constructor() {
-    this.chatgpt = new ChatGPTAgent();
+  constructor(logger: Logger) {
+    this.chatgpt = new ChatGPTAgent(logger);
   }
 
   /**

src/eval/evaluators/single-attempt.ts

@@ -25,6 +25,7 @@
  * ```
  */
 
+import type { Logger } from '../../utils/logger.js';
 import { attemptMetricsSchema } from '../core/schemas.js';
 import type { AttemptMetrics } from '../core/types.js';
 import { ChatGPTAgent } from './chatgpt-agent.js';
@@ -54,9 +55,11 @@ export class SingleAttemptEvaluator {
 
   /**
    * Create a new single-attempt evaluator
+   *
+   * @param logger - Logger for progress messages (reserved for future use)
    */
-  constructor() {
-    this.chatgpt = new ChatGPTAgent();
+  constructor(logger: Logger) {
+    this.chatgpt = new ChatGPTAgent(logger);
   }
 
   /**