From 91f4c4299df8b906e45cafd4d82418fea80b272e Mon Sep 17 00:00:00 2001
From: yhAutomationQA <yhautomationqa@gmail.com>
Date: Sun, 10 May 2026 22:31:58 -0400
Subject: [PATCH 1/9] Add AI PR review system

---
 .ai/prompts/pr-review.prompt.md          |  45 +++
 .ai/skills/playwright-pr-review.md       | 128 +++++++
 .github/PULL_REQUEST_TEMPLATE.md         |  78 +++++
 .github/scripts/ai-review.mjs            | 220 ++++++++++++
 .github/workflows/opencode-pr-review.yml | 425 +++++++++++++++++++++++
 5 files changed, 896 insertions(+)
 create mode 100644 .ai/prompts/pr-review.prompt.md
 create mode 100644 .ai/skills/playwright-pr-review.md
 create mode 100644 .github/PULL_REQUEST_TEMPLATE.md
 create mode 100644 .github/scripts/ai-review.mjs
 create mode 100644 .github/workflows/opencode-pr-review.yml

diff --git a/.ai/prompts/pr-review.prompt.md b/.ai/prompts/pr-review.prompt.md
new file mode 100644
index 0000000..1ed555a
--- /dev/null
+++ b/.ai/prompts/pr-review.prompt.md
@@ -0,0 +1,45 @@
+# AI PR Review Prompt — Playwright + TypeScript
+
+You are reviewing a pull request for a **Playwright + TypeScript test automation framework** following enterprise QA engineering best practices.
+
+## Review Context
+
+- **Framework:** Playwright + TypeScript (v1.59+)
+- **Test Runner:** `@playwright/test`
+- **Pattern:** Page Object Model with `BasePage` abstraction
+- **Fixtures:** Dependency injection via `fixtures/index.ts`
+- **Config:** Multi-environment via `config/env-manager.ts` and `.env.*` files
+- **Logger:** Pino-based structured logging (`utils/logger.ts`)
+- **Test Data:** Centralized in `utils/test-data.ts`
+- **API Layer:** `ApiClient` helper wrapping Playwright's `APIRequestContext`
+- **Locators:** CSS selectors and `data-test` attributes; defined as `private readonly` in page objects
+- **Auth:** `storageState` pattern in `mcp/auth-reuse.ts`; per-test login via fixtures
+- **CI:** Fully parallel execution, retries on CI, one worker per browser in CI
+
+## Load the Skill
+
+First, load the `.ai/skills/playwright-pr-review.md` skill file. All rules, severity levels, and output formatting from that skill must be followed.
+
+## Task
+
+1. Read the git diff below (between PR branch and `main`).
+2. Analyze all changed files against the review checklist in the skill file.
+3. Ignore files that are not related to the Playwright/TypeScript framework (e.g., documentation-only changes, CI config changes, dependency bumps).
+4. For each issue found, output a severity-labeled block.
+5. End with the PR Quality Score summary table.
+
+## Git Diff
+
+```
+{{GIT_DIFF}}
+```
+
+## Review Guidelines
+
+- Focus on what the diff **introduces** — not on existing code outside the diff.
+- If a change violates a framework pattern, note which pattern it breaks and reference the correct convention.
+- For anti-patterns, provide the **exact** recommended fix as runnable code.
+- Do NOT suggest auto-merge or approve the PR. The review requires human approval.
+- If no issues are found, output: **"No issues found. PR adheres to framework standards."** with a score of 100.
+
+Remember: prioritize stability over speed. Flag anything that could cause CI flakiness, even if it seems minor.
diff --git a/.ai/skills/playwright-pr-review.md b/.ai/skills/playwright-pr-review.md
new file mode 100644
index 0000000..4bfab92
--- /dev/null
+++ b/.ai/skills/playwright-pr-review.md
@@ -0,0 +1,128 @@
+# Playwright + TypeScript PR Review Skill
+
+## Role
+
+You are a **Senior QA Automation Architect** with deep expertise in Playwright, TypeScript, and enterprise test automation frameworks. You review pull requests with a focus on stability, maintainability, and CI reliability.
+
+## Scope
+
+Analyze ONLY the git diff — the lines added, modified, or removed. Do not review unchanged code unless context is required to understand the change.
+
+## Severity Levels
+
+| Level     | Meaning                                                                 | Action Required        |
+|-----------|-------------------------------------------------------------------------|------------------------|
+| Critical  | Causes CI flakiness, test false-positive/false-negative, or data loss   | Must fix before merge  |
+| Major     | Violates framework patterns, DRY, or will degrade stability over time   | Should fix before merge|
+| Minor     | Code style, readability, or edge cases                                  | Consider fixing        |
+| Suggestion| Improvement idea, not a defect                                          | Optional               |
+
+## Review Checklist
+
+### Flaky Test Patterns
+- [ ] Hardcoded waits (`page.waitForTimeout`) — **Critical** — use `waitForSelector`, `waitForURL`, `waitForResponse`, or `toBeVisible` assertions instead
+- [ ] Race conditions — missing `waitForLoadState('networkidle')` after `page.goto`; navigation-dependent actions without proper waits
+- [ ] Element detached between query and action — accessing locator properties without re-query
+- [ ] `Promise.all` with competing clicks/navigations without proper ordering
+- [ ] `.catch(() => {})` or `.catch(() => null)` — silent error swallowing masks real failures
+- [ ] Tests depending on test execution order (`test.describe.serial`, shared mutable state)
+- [ ] Missing `await` on Playwright actions — returns a `Locator` or `Promise` instead of executing
+
+### Improper Locator Usage
+- [ ] CSS selectors tied to unstable classes (e.g., `btn_primary_3f6b4`, hash-suffixed classes)
+- [ ] Chained CSS (`div > ul > li > a`) — brittle against DOM structure changes
+- [ ] `page.$` / `page.$$` (legacy API) — use `page.locator()` instead
+- [ ] `page.waitForSelector` — use `locator.waitFor()` for consistency
+- [ ] Locators that match multiple elements when one is expected, without `.first()` or `.nth()`
+- [ ] Missing `data-testid` / `data-test` attributes — prefer `page.getByTestId()` over CSS
+- [ ] Text-based selectors without `exact` matching when ambiguity exists
+- [ ] XPath when CSS or getByRole would be more stable
+
+### Missing Assertions
+- [ ] Navigation actions without URL assertions (`await expect(page).toHaveURL(...)`)
+- [ ] Click actions without state verification (what should appear/disappear?)
+- [ ] API responses without status code or body validation
+- [ ] Form submissions without success/error state assertion
+- [ ] Asynchronous operations without waiting for completion
+
+### Anti-Patterns in Playwright
+- [ ] `page.evaluate` for reading element properties — use locator assertions instead
+- [ ] `page.evaluate` for DOM manipulation — use Playwright actions (`click`, `fill`)
+- [ ] Mixing `@playwright/test` and raw `playwright` APIs in tests
+- [ ] Direct `browser.newPage()` in tests instead of using the `page` fixture
+- [ ] Manual cleanup (`browser.close()`) in tests — fixtures handle teardown
+- [ ] Overusing `{ force: true }` — bypasses actionability checks, creates false passes
+- [ ] `page.screenshot` without `fullPage: true` when capturing full content
+- [ ] Modifying `page.on('dialog')` handlers without cleanup (`page.off`)
+- [ ] Ignoring `page.context()` and `page.request()` for API integration tests
+
+### TypeScript Issues
+- [ ] `any` type usage — prefer specific types or `unknown` with narrowing
+- [ ] `as` casts that suppress real type mismatches (`as never`, `as any`)
+- [ ] Missing return type annotations on exported functions
+- [ ] Implicit `any` in callback parameters
+- [ ] Incorrect or missing generic type parameters
+- [ ] Importing types as values or values as types
+- [ ] Unused imports or variables (detectable by the linter)
+- [ ] Async function without `await` on promise-returning calls
+
+### Framework Architecture Violations
+- [ ] Business logic or assertions in page objects — page objects should expose state, not assert
+- [ ] Tests that bypass the fixture system — creating page objects manually in test bodies
+- [ ] Page objects using `page` directly instead of `BasePage` wrappers
+- [ ] Missing page object for a new page/section of the application
+- [ ] Configuration hardcoded in tests instead of using `config/env-manager.ts`
+- [ ] Environment-specific values in test code instead of `.env.*` files
+- [ ] Data setup via UI when API setup is available (`fixtures/index.ts` pattern)
+- [ ] Breaking the `test → page object → helper` layering
+
+### DRY Violations
+- [ ] Login flow duplicated across files — use `LoginPage` page object + fixture
+- [ ] Locator selectors duplicated as strings — define once in page objects
+- [ ] Repeated assertion patterns — extract to custom assertions or helpers
+- [ ] Navigation logic duplicated — centralize in page objects
+- [ ] Test data duplicated — use `test-data.ts` exports
+- [ ] API call patterns duplicated — use `ApiClient` helper
+- [ ] Common setup/teardown logic in individual tests instead of hooks or fixtures
+
+## Output Format
+
+Every review comment must include:
+
+```markdown
+## Severity: <Critical | Major | Minor | Suggestion>
+**File:** `<file-path>:<line-number>`
+**Issue:** <one-line summary>
+**Explanation:** <why this is a problem, with specific reasoning>
+**Recommended Fix:** <code snippet or actionable instruction>
+```
+
+## PR Quality Score
+
+At the end of the review, compute a score:
+
+- Start at **100**
+- **-15** per Critical
+- **-10** per Major
+- **-5** per Minor
+- **-0** per Suggestion
+
+| Score    | Rating       |
+|----------|--------------|
+| 100      | Excellent    |
+| 80-99    | Good         |
+| 60-79    | Needs Work   |
+| <60      | Failing      |
+
+Include the score as a summary table:
+
+```markdown
+## PR Quality Score: <N>/100 — <Rating>
+
+| Severity   | Count |
+|------------|-------|
+| Critical   | N     |
+| Major      | N     |
+| Minor      | N     |
+| Suggestion | N     |
+```
diff --git a/.github/PULL_REQUEST_TEMPLATE.md b/.github/PULL_REQUEST_TEMPLATE.md
new file mode 100644
index 0000000..f3b4c9e
--- /dev/null
+++ b/.github/PULL_REQUEST_TEMPLATE.md
@@ -0,0 +1,78 @@
+# Pull Request — Playwright TypeScript Framework
+
+## Description
+
+<!-- Provide a clear, concise description of the changes in this PR.
+     Include the motivation for the change and how it improves the framework. -->
+
+## Type of Change
+
+<!-- Check all that apply. -->
+
+- [ ] New test(s)
+- [ ] Test fix / flakiness resolution
+- [ ] Framework enhancement
+- [ ] Bug fix (test or framework)
+- [ ] Refactoring / code quality
+- [ ] Documentation
+- [ ] CI / DevOps
+- [ ] Dependency update
+- [ ] Other (describe below)
+
+## Affected Areas
+
+<!-- List the pages, components, or features affected by this change. -->
+
+- 
+
+## Test Evidence
+
+<!-- Required for all non-documentation PRs. Provide evidence that the change works correctly. -->
+
+- [ ] All existing tests pass locally: `npx playwright test`
+- [ ] New tests pass: `npx playwright test <spec-file>`
+- [ ] Tests pass across browsers: `npx playwright test --project=chromium --project=firefox --project=webkit`
+- [ ] Trace/screenshot artifacts reviewed for correctness
+- [ ] Verified against the target environment: `dev` / `qa` / `staging` / `prod`
+
+## Self-Review Checklist
+
+<!-- Confirm each item before requesting review. -->
+
+### Flakiness Prevention
+- [ ] No hardcoded waits (`page.waitForTimeout`) — used `waitForSelector`, `waitForURL`, or assertions
+- [ ] No `.catch(() => {})` or silent error swallowing
+- [ ] All Playwright actions have proper `await`
+- [ ] Navigation actions are followed by URL assertions or element waits
+- [ ] Locators use stable selectors (`data-test` attributes preferred)
+- [ ] No dependency on test execution order
+
+### Framework Compliance
+- [ ] Tests use fixtures from `fixtures/index.ts` — no manual page object instantiation
+- [ ] Page objects extend `BasePage` and define locators as `private readonly`
+- [ ] No UI logic or assertions inside page objects
+- [ ] Test data sourced from `utils/test-data.ts` or `config/env-manager.ts`
+- [ ] No hardcoded URLs, credentials, or environment-specific values
+- [ ] Follows existing coding conventions (naming, imports, formatting)
+
+### Quality
+- [ ] No `any` types — used specific types or `unknown` with narrowing
+- [ ] No `as` casts that suppress type mismatches
+- [ ] No duplicate code — reused existing page objects, fixtures, and helpers
+- [ ] Linter passes: `npm run lint`
+- [ ] TypeScript compiles: `npx tsc --noEmit`
+
+## PR Quality Score (self-assessed)
+
+<!-- Compute based on the checklist above. -->
+
+| Severity   | Count |
+|------------|-------|
+| Blocking   | N     |
+| Major      | N     |
+| Minor      | N     |
+| Clean      | N     |
+
+---
+
+> **Note:** Human approval is required before merge. This repository enforces branch protection rules — all checks must pass and a maintainer must approve.
diff --git a/.github/scripts/ai-review.mjs b/.github/scripts/ai-review.mjs
new file mode 100644
index 0000000..8e867db
--- /dev/null
+++ b/.github/scripts/ai-review.mjs
@@ -0,0 +1,220 @@
+/**
+ * ai-review.mjs — AI-powered PR review script
+ *
+ * This script is the AI provider fallback for the opencode-pr-review workflow.
+ * It reads the skill definition, the prompt template, and the git diff,
+ * then sends them to an AI API (OpenAI or Anthropic) for analysis.
+ *
+ * Usage:
+ *   node .github/scripts/ai-review.mjs \
+ *     --provider openai \
+ *     --api-key sk-... \
+ *     --model gpt-4o \
+ *     --skill-file .ai/skills/playwright-pr-review.md \
+ *     --prompt-file .ai/prompts/pr-review.prompt.md \
+ *     --diff-file /tmp/pr-diff.txt
+ *
+ * Environment variables (alternative to CLI args):
+ *   AI_REVIEW_PROVIDER, AI_REVIEW_API_KEY, AI_REVIEW_MODEL
+ */
+
+import fs from 'node:fs';
+import path from 'node:path';
+
+// ---------------------------------------------------------------------------
+// CLI argument parsing
+// ---------------------------------------------------------------------------
+function parseArgs() {
+  const args = process.argv.slice(2);
+  const opts = {};
+
+  for (let i = 0; i < args.length; i++) {
+    switch (args[i]) {
+      case '--provider':
+        opts.provider = args[++i];
+        break;
+      case '--api-key':
+        opts.apiKey = args[++i];
+        break;
+      case '--model':
+        opts.model = args[++i];
+        break;
+      case '--skill-file':
+        opts.skillFile = args[++i];
+        break;
+      case '--prompt-file':
+        opts.promptFile = args[++i];
+        break;
+      case '--diff-file':
+        opts.diffFile = args[++i];
+        break;
+      default:
+        // skip
+    }
+  }
+
+  // Resolve with environment variable fallbacks
+  opts.provider = opts.provider || process.env.AI_REVIEW_PROVIDER || 'openai';
+  opts.apiKey = opts.apiKey || process.env.AI_REVIEW_API_KEY || '';
+  opts.model = opts.model || process.env.AI_REVIEW_MODEL || 'gpt-4o';
+
+  return opts;
+}
+
+// ---------------------------------------------------------------------------
+// File reading helpers
+// ---------------------------------------------------------------------------
+function readFileOrExit(filePath, label) {
+  const resolved = path.resolve(filePath);
+  try {
+    return fs.readFileSync(resolved, 'utf-8');
+  } catch (err) {
+    console.error(`Error reading ${label} at ${resolved}: ${err.message}`);
+    process.exit(1);
+  }
+}
+
+// ---------------------------------------------------------------------------
+// Prompt builder — injects the diff into the prompt template
+// ---------------------------------------------------------------------------
+function buildPrompt(skillContent, promptTemplate, diffContent) {
+  // Prepend the skill rules to the prompt so the AI has full context
+  const systemPrompt = `You are a Senior QA Automation Architect. Follow these rules:\n\n${skillContent}`;
+
+  // Inject the diff into the prompt template
+  const userPrompt = promptTemplate.replace('{{GIT_DIFF}}', diffContent);
+
+  return { systemPrompt, userPrompt };
+}
+
+// ---------------------------------------------------------------------------
+// AI API callers
+// ---------------------------------------------------------------------------
+async function callOpenAI(apiKey, model, systemPrompt, userPrompt) {
+  const response = await fetch('https://api.openai.com/v1/chat/completions', {
+    method: 'POST',
+    headers: {
+      'Content-Type': 'application/json',
+      Authorization: `Bearer ${apiKey}`,
+    },
+    body: JSON.stringify({
+      model: model,
+      messages: [
+        { role: 'system', content: systemPrompt },
+        { role: 'user', content: userPrompt },
+      ],
+      temperature: 0.1,       // low temperature for consistent, deterministic review
+      max_tokens: 4096,       // enough for a detailed review
+    }),
+  });
+
+  if (!response.ok) {
+    const errorBody = await response.text();
+    throw new Error(`OpenAI API error ${response.status}: ${errorBody}`);
+  }
+
+  const data = await response.json();
+  return data.choices[0].message.content;
+}
+
+async function callAnthropic(apiKey, model, systemPrompt, userPrompt) {
+  const response = await fetch('https://api.anthropic.com/v1/messages', {
+    method: 'POST',
+    headers: {
+      'Content-Type': 'application/json',
+      'x-api-key': apiKey,
+      'anthropic-version': '2023-06-01',
+    },
+    body: JSON.stringify({
+      model: model,
+      system: systemPrompt,
+      messages: [
+        { role: 'user', content: userPrompt },
+      ],
+      temperature: 0.1,
+      max_tokens: 4096,
+    }),
+  });
+
+  if (!response.ok) {
+    const errorBody = await response.text();
+    throw new Error(`Anthropic API error ${response.status}: ${errorBody}`);
+  }
+
+  const data = await response.json();
+  return data.content[0].text;
+}
+
+// ---------------------------------------------------------------------------
+// Main
+// ---------------------------------------------------------------------------
+async function main() {
+  const opts = parseArgs();
+
+  // Validate required inputs
+  if (!opts.diffFile || !opts.promptFile || !opts.skillFile) {
+    console.error(
+      'Usage: node ai-review.mjs --provider <provider> --api-key <key> --model <model> ' +
+      '--skill-file <path> --prompt-file <path> --diff-file <path>',
+    );
+    process.exit(1);
+  }
+
+  // Verify the diff actually has content
+  const diffContent = readFileOrExit(opts.diffFile, 'diff file');
+  if (!diffContent.trim() || diffContent.trim() === '(no diff — no Playwright/TypeScript files changed)') {
+    console.log('## AI PR Review');
+    console.log('');
+    console.log('**No issues found.** No Playwright/TypeScript code changes detected in this diff.');
+    console.log('');
+    console.log('## PR Quality Score: 100/100 — Excellent');
+    console.log('');
+    console.log('| Severity | Count |');
+    console.log('|----------|-------|');
+    console.log('| Critical | 0 |');
+    console.log('| Major | 0 |');
+    console.log('| Minor | 0 |');
+    console.log('| Suggestion | 0 |');
+    return;
+  }
+
+  const skillContent = readFileOrExit(opts.skillFile, 'skill file');
+  const promptTemplate = readFileOrExit(opts.promptFile, 'prompt template');
+
+  const { systemPrompt, userPrompt } = buildPrompt(skillContent, promptTemplate, diffContent);
+
+  if (!opts.apiKey) {
+    console.error('Error: No API key provided. Set AI_REVIEW_API_KEY environment variable or pass --api-key.');
+    process.exit(1);
+  }
+
+  // Call the appropriate AI provider
+  let reviewText;
+  switch (opts.provider) {
+    case 'openai':
+      reviewText = await callOpenAI(opts.apiKey, opts.model, systemPrompt, userPrompt);
+      break;
+    case 'anthropic':
+      reviewText = await callAnthropic(opts.apiKey, opts.model, systemPrompt, userPrompt);
+      break;
+    default:
+      console.error(`Unsupported provider: ${opts.provider}. Use 'openai' or 'anthropic'.`);
+      process.exit(1);
+  }
+
+  // Print the review to stdout (captured by the workflow)
+  console.log(reviewText);
+}
+
+main().catch((err) => {
+  console.error('AI review script failed:', err.message);
+  // Output a fallback review so the workflow doesn't break
+  console.log('## AI PR Review');
+  console.log('');
+  console.log('**Review could not be completed.**');
+  console.log('');
+  console.log(`Error: ${err.message}`);
+  console.log('');
+  console.log('## PR Quality Score: N/A');
+  process.exit(1);
+});
diff --git a/.github/workflows/opencode-pr-review.yml b/.github/workflows/opencode-pr-review.yml
new file mode 100644
index 0000000..5d18e88
--- /dev/null
+++ b/.github/workflows/opencode-pr-review.yml
@@ -0,0 +1,425 @@
+# =============================================================================
+# opencode-pr-review.yml — AI-Powered PR Review for Playwright + TypeScript
+# =============================================================================
+#
+# This workflow triggers on pull requests to `main`, analyzes the git diff
+# using an AI agent (OpenCode or OpenAI-compatible API), and posts a
+# structured code review with severity-labeled findings and a quality score.
+#
+# Design Principles:
+# - Modular: review logic is in a reusable action; the workflow orchestrates
+# - Observable: all review output is captured as both PR comment + check run
+# - Non-blocking for merge — human approval is required (branch protection)
+# - Idempotent: re-running replaces the previous review comment
+# - Cost-aware: only reviews the diff (not the full file) to minimize tokens
+#
+# Branch Protection Compatibility:
+# This workflow creates a check run named "AI PR Review / review". You can
+# require this check to pass in GitHub branch protection rules.
+#
+# Security:
+# - AI API keys stored as GitHub Secrets (never in code or logs)
+# - Pull request comments posted by github-actions bot — no custom tokens
+# - Works with existing branch protection without requiring auto-merge
+# =============================================================================
+
+name: AI PR Review
+
+on:
+  # Trigger on PRs to main — includes draft and ready-for-review
+  pull_request:
+    branches: [main, master]
+    types: [opened, synchronize, reopened, ready_for_review]
+
+  # Allow manual trigger for ad-hoc reviews
+  # workflow_dispatch:
+  #   inputs:
+  #     pr_number:
+  #       description: Pull request number to review
+  #       required: true
+
+# Ensure only one review runs per PR at a time to avoid race conditions
+# on the PR comment. New commits cancel in-progress reviews.
+concurrency:
+  group: pr-review-${{ github.event.pull_request.number || github.event.inputs.pr_number }}
+  cancel-in-progress: true
+
+env:
+  # Node.js version — must match the framework
+  NODE_VERSION: 22
+
+jobs:
+  # ===========================================================================
+  # Job 1: review
+  #   - Checks out the PR branch
+  #   - Computes the git diff against the base branch
+  #   - Sends the diff to an AI agent for analysis
+  #   - Posts structured review as a PR comment
+  # ===========================================================================
+  review:
+    # Skip for draft PRs (optional — remove if you want draft reviews)
+    if: github.event.pull_request.draft == false
+
+    runs-on: ubuntu-latest
+    timeout-minutes: 15
+
+    # Grant permissions needed to post PR comments and create check runs
+    permissions:
+      contents: read
+      pull-requests: write
+      checks: write
+
+    steps:
+      # -----------------------------------------------------------------------
+      # Step 1: Validate that the AI provider is configured
+      # -----------------------------------------------------------------------
+      - name: Validate AI provider configuration
+        id: validate
+        run: |
+          PROVIDER="${{ secrets.AI_REVIEW_PROVIDER || 'openai' }}"
+          echo "provider=$PROVIDER" >> $GITHUB_OUTPUT
+
+          case "$PROVIDER" in
+            openai)
+              if [ -z "${{ secrets.OPENAI_API_KEY }}" ]; then
+                echo "Error: OPENAI_API_KEY secret is not set."
+                echo "Set it in your repository secrets: Settings > Secrets and variables > Actions"
+                exit 1
+              fi
+              ;;
+            anthropic)
+              if [ -z "${{ secrets.ANTHROPIC_API_KEY }}" ]; then
+                echo "Error: ANTHROPIC_API_KEY secret is not set."
+                exit 1
+              fi
+              ;;
+            opencode)
+              if [ -z "${{ secrets.OPENCODE_API_KEY }}" ]; then
+                echo "Warning: OPENCODE_API_KEY not set — using opencode CLI default."
+              fi
+              ;;
+            *)
+              echo "Error: Unknown provider '$PROVIDER'. Supported: openai, anthropic, opencode"
+              exit 1
+              ;;
+          esac
+
+      # -----------------------------------------------------------------------
+      # Step 2: Check out the PR branch with full git history
+      # -----------------------------------------------------------------------
+      - name: Checkout PR branch
+        uses: actions/checkout@v4
+        with:
+          # Fetch full history so we can diff against the base branch
+          fetch-depth: 0
+          # Use the merge commit to get the correct diff against base
+          ref: ${{ github.event.pull_request.head.sha }}
+
+      # -----------------------------------------------------------------------
+      # Step 3: Compute the git diff
+      # -----------------------------------------------------------------------
+      - name: Compute git diff against base branch
+        id: diff
+        run: |
+          BASE_SHA="${{ github.event.pull_request.base.sha }}"
+          HEAD_SHA="${{ github.event.pull_request.head.sha }}"
+
+          echo "Base SHA: $BASE_SHA"
+          echo "Head SHA: $HEAD_SHA"
+
+          # Get the diff — only the lines changed, with context
+          DIFF=$(git diff "$BASE_SHA"..."$HEAD_SHA" -- \
+            ':!package-lock.json' \
+            ':!*.lock' \
+            ':!*.png' \
+            ':!*.jpg' \
+            ':!*.jpeg' \
+            ':!*.gif' \
+            ':!*.svg' \
+            ':!*.ico' \
+            ':!*.woff*' \
+            ':!*.eot' \
+            ':!*.ttf' \
+            2>/dev/null || echo "(empty diff or error)")
+
+          # Handle empty or large diffs
+          if [ -z "$DIFF" ]; then
+            echo "diff_empty=true" >> $GITHUB_OUTPUT
+            echo "diff_content=(no diff — no Playwright/TypeScript files changed)" >> $GITHUB_OUTPUT
+          else
+            # Truncate diff to prevent token overflow (conservative limit)
+            MAX_DIFF_LENGTH=$(( 4 * 1024 * 50 ))  # ~50K tokens worth of diff
+            if [ ${#DIFF} -gt $MAX_DIFF_LENGTH ]; then
+              DIFF="${DIFF:0:$MAX_DIFF_LENGTH}"
+              DIFF+="\n\n[... diff truncated at ${MAX_DIFF_LENGTH} characters — only the first part was analyzed ...]"
+            fi
+            echo "diff_empty=false" >> $GITHUB_OUTPUT
+            echo "diff_content<<EOF_DIFF" >> $GITHUB_OUTPUT
+            echo "$DIFF" >> $GITHUB_OUTPUT
+            echo "EOF_DIFF" >> $GITHUB_OUTPUT
+          fi
+
+      # -----------------------------------------------------------------------
+      # Step 4: Install Node.js and dependencies
+      # -----------------------------------------------------------------------
+      - name: Setup Node.js
+        uses: actions/setup-node@v4
+        with:
+          node-version: ${{ env.NODE_VERSION }}
+          cache: npm
+
+      - name: Install dependencies
+        run: npm ci --ignore-scripts
+        env:
+          # Skip Playwright browser install — not needed for review
+          PLAYWRIGHT_SKIP_BROWSER_DOWNLOAD: 1
+
+      # -----------------------------------------------------------------------
+      # Step 5 (Option A): Run AI review via OpenCode CLI
+      #     Uses the OpenCode CLI tool to run the PR review skill against the
+      #     prompt template. Falls back to API-based review if not configured.
+      # -----------------------------------------------------------------------
+      - name: Run AI review (OpenCode)
+        id: review
+        run: |
+          # Write diff to a temp file so it can be referenced by the prompt
+          echo "DIFF_EMPTY=${{ steps.diff.outputs.diff_empty }}" >> $GITHUB_ENV
+
+          if [ "${{ steps.diff.outputs.diff_empty }}" = "true" ]; then
+            echo "review_output=No issues found. No Playwright/TypeScript code changed." >> $GITHUB_OUTPUT
+            echo "review_score=100" >> $GITHUB_OUTPUT
+            echo "review_rating=Excellent" >> $GITHUB_OUTPUT
+            exit 0
+          fi
+
+          # Write the diff to a temporary file for the review script
+          echo "${{ steps.diff.outputs.diff_content }}" > /tmp/pr-diff.txt
+
+          # Check for opencode CLI
+          if command -v npx &> /dev/null && npx --yes opencode --help &> /dev/null 2>&1; then
+            # Run the review using opencode with the skill and prompt
+            # This approach loads the skill, reads the prompt template, injects the diff
+            echo "Running review via OpenCode..."
+            npx opencode review \
+              --skill .ai/skills/playwright-pr-review.md \
+              --prompt .ai/prompts/pr-review.prompt.md \
+              --diff /tmp/pr-diff.txt \
+              --output /tmp/review-output.txt \
+              --format markdown \
+              2>&1 || true
+
+            if [ -f /tmp/review-output.txt ]; then
+              REVIEW=$(cat /tmp/review-output.txt)
+            else
+              REVIEW="(OpenCode review produced no output)"  
+            fi
+          else
+            # If opencode CLI is not available, try using an AI API directly
+            echo "OpenCode CLI not detected. Using AI API fallback..."
+            REVIEW=$(node .github/scripts/ai-review.mjs \
+              --provider "${{ steps.validate.outputs.provider }}" \
+              --api-key "${{ secrets.OPENAI_API_KEY || secrets.ANTHROPIC_API_KEY || '' }}" \
+              --model "${{ secrets.AI_REVIEW_MODEL || 'gpt-4o' }}" \
+              --skill-file .ai/skills/playwright-pr-review.md \
+              --prompt-file .ai/prompts/pr-review.prompt.md \
+              --diff-file /tmp/pr-diff.txt \
+              2>&1 || echo "(AI review script failed — see logs)")
+          fi
+
+          # Write review output
+          echo "review_output<<EOF_REVIEW" >> $GITHUB_OUTPUT
+          echo "$REVIEW" >> $GITHUB_OUTPUT
+          echo "EOF_REVIEW" >> $GITHUB_OUTPUT
+
+          # Extract score from review output (format: "## PR Quality Score: <N>/100")
+          SCORE=$(echo "$REVIEW" | grep -oP 'PR Quality Score:\s*\K(\d+)' || echo "")
+          if [ -z "$SCORE" ]; then
+            # Fallback: parse severity counts to compute score
+            CRIT=$(echo "$REVIEW" | grep -c '^|.*Critical.*|' || echo 0)
+            MAJ=$(echo "$REVIEW" | grep -c '^|.*Major.*|' || echo 0)
+            MIN=$(echo "$REVIEW" | grep -c '^|.*Minor.*|' || echo 0)
+            SCORE=$(( 100 - (CRIT * 15) - (MAJ * 10) - (MIN * 5) ))
+            [ "$SCORE" -lt 0 ] && SCORE=0
+          fi
+          echo "review_score=$SCORE" >> $GITHUB_OUTPUT
+
+          # Compute rating
+          if [ "$SCORE" -eq 100 ]; then RATING="Excellent"
+          elif [ "$SCORE" -ge 80 ]; then RATING="Good"
+          elif [ "$SCORE" -ge 60 ]; then RATING="Needs Work"
+          else RATING="Failing"
+          fi
+          echo "review_rating=$RATING" >> $GITHUB_OUTPUT
+
+      # -----------------------------------------------------------------------
+      # Step 6: Post review as a PR comment
+      #     Replaces any previous AI review comment to keep the PR clean.
+      # -----------------------------------------------------------------------
+      - name: Post PR review comment
+        uses: actions/github-script@v7
+        id: comment
+        with:
+          script: |
+            const { data: comments } = await github.rest.issues.listComments({
+              owner: context.repo.owner,
+              repo: context.repo.repo,
+              issue_number: context.issue.number,
+            });
+
+            // Find and delete any previous AI review comment
+            const marker = '<!-- opencode-pr-review -->';
+            for (const comment of comments) {
+              if (comment.body && comment.body.includes(marker)) {
+                await github.rest.issues.deleteComment({
+                  owner: context.repo.owner,
+                  repo: context.repo.repo,
+                  comment_id: comment.id,
+                });
+              }
+            }
+
+            // Build the review body
+            const reviewOutput = `${{ steps.review.outputs.review_output }}`;
+            const score = `${{ steps.review.outputs.review_score }}`;
+            const rating = `${{ steps.review.outputs.review_rating }}`;
+
+            const header = `## 🤖 AI PR Review\n\n`;
+            const footer = `\n\n---\n*This review was generated automatically by the AI PR Review workflow. Human approval is required before merge.*\n${marker}`;
+
+            let body;
+            if (reviewOutput.includes('No issues found')) {
+              body = header + `**No issues found.** PR adheres to framework standards.\n\n**PR Quality Score:** ${score}/100 — ${rating}` + footer;
+            } else {
+              body = header + reviewOutput + footer;
+            }
+
+            // Post the new review comment
+            await github.rest.issues.createComment({
+              owner: context.repo.owner,
+              repo: context.repo.repo,
+              issue_number: context.issue.number,
+              body: body,
+            });
+
+            return score;
+
+      # -----------------------------------------------------------------------
+      # Step 7: Create a check run for branch protection compatibility
+      #     This check can be marked as "required" in branch protection rules.
+      #     The check passes if score >= 60, fails if < 60.
+      # -----------------------------------------------------------------------
+      - name: Create check run
+        uses: actions/github-script@v7
+        with:
+          script: |
+            const score = parseInt('${{ steps.review.outputs.review_score }}' || '100', 10);
+            const rating = '${{ steps.review.outputs.review_rating }}' || 'Good';
+            const headSha = '${{ github.event.pull_request.head.sha }}';
+
+            // Determine conclusion based on score
+            let conclusion = 'neutral';
+            let summary = `Score: ${score}/100 — ${rating}`;
+            let text = '';
+
+            if (score >= 80) {
+              conclusion = 'success';
+              text = 'PR meets quality standards. No blocking issues found.';
+            } else if (score >= 60) {
+              conclusion = 'neutral';
+              text = 'PR has some issues that should be addressed. Review the comment above.';
+            } else {
+              conclusion = 'failure';
+              text = 'PR has critical issues that must be resolved before merge.';
+            }
+
+            await github.rest.checks.create({
+              owner: context.repo.owner,
+              repo: context.repo.repo,
+              name: 'AI PR Review / review',
+              head_sha: headSha,
+              status: 'completed',
+              conclusion: conclusion,
+              output: {
+                title: `AI PR Review: ${score}/100 — ${rating}`,
+                summary: summary,
+                text: text + '\n\nSee the PR comment for detailed findings.',
+              },
+            });
+
+      # -----------------------------------------------------------------------
+      # Step 8: Update PR labels based on quality score
+      # -----------------------------------------------------------------------
+      - name: Update PR labels
+        uses: actions/github-script@v7
+        with:
+          script: |
+            const score = parseInt('${{ steps.review.outputs.review_score }}' || '100', 10);
+            const labels = new Set();
+
+            // Remove existing AI review labels
+            const { data: currentLabels } = await github.rest.issues.listLabelsOnIssue({
+              owner: context.repo.owner,
+              repo: context.repo.repo,
+              issue_number: context.issue.number,
+            });
+
+            const aiLabels = ['ai-review-passed', 'ai-review-failed', 'ai-review-warning'];
+            const toRemove = currentLabels
+              .filter(l => aiLabels.includes(l.name))
+              .map(l => l.name);
+
+            for (const label of toRemove) {
+              try {
+                await github.rest.issues.removeLabel({
+                  owner: context.repo.owner,
+                  repo: context.repo.repo,
+                  issue_number: context.issue.number,
+                  name: label,
+                });
+              } catch { /* label may not exist */ }
+            }
+
+            // Add the appropriate label
+            let newLabel;
+            if (score >= 80) {
+              newLabel = 'ai-review-passed';
+            } else if (score >= 60) {
+              newLabel = 'ai-review-warning';
+            } else {
+              newLabel = 'ai-review-failed';
+            }
+
+            try {
+              await github.rest.issues.addLabels({
+                owner: context.repo.owner,
+                repo: context.repo.repo,
+                issue_number: context.issue.number,
+                labels: [newLabel],
+              });
+            } catch (err) {
+              // Label may not exist in the repository — create it
+              console.log('Could not add label. Create it in your repository settings.');
+            }
+
+  # ===========================================================================
+  # Job 2: summary-report (optional)
+  #   Posts a summary to the workflow run for easy visibility.
+  # ===========================================================================
+  summary-report:
+    needs: review
+    if: always()
+    runs-on: ubuntu-latest
+    steps:
+      - name: Post workflow summary
+        run: |
+          echo "## AI PR Review Summary" >> $GITHUB_STEP_SUMMARY
+          echo "" >> $GITHUB_STEP_SUMMARY
+          echo "| Metric | Value |" >> $GITHUB_STEP_SUMMARY
+          echo "|--------|-------|" >> $GITHUB_STEP_SUMMARY
+          echo "| PR | #${{ github.event.pull_request.number }} |" >> $GITHUB_STEP_SUMMARY
+          echo "| Branch | ${{ github.event.pull_request.head.ref }} → ${{ github.event.pull_request.base.ref }} |" >> $GITHUB_STEP_SUMMARY
+          echo "| Author | ${{ github.event.pull_request.user.login }} |" >> $GITHUB_STEP_SUMMARY
+          echo "| Quality Score | ${{ needs.review.outputs.review_score || 'N/A' }} / 100 |" >> $GITHUB_STEP_SUMMARY
+          echo "| Rating | ${{ needs.review.outputs.review_rating || 'N/A' }} |" >> $GITHUB_STEP_SUMMARY
+          echo "" >> $GITHUB_STEP_SUMMARY
+          echo "> **Note:** This is an automated review. A human maintainer must approve the PR before merge." >> $GITHUB_STEP_SUMMARY

From d97f94824d7b09579be324d03937949e0f23529b Mon Sep 17 00:00:00 2001
From: yhAutomationQA <yhautomationqa@gmail.com>
Date: Sun, 10 May 2026 22:34:51 -0400
Subject: [PATCH 2/9] align label names with repo

---
 .github/workflows/opencode-pr-review.yml | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/.github/workflows/opencode-pr-review.yml b/.github/workflows/opencode-pr-review.yml
index 5d18e88..869bf12 100644
--- a/.github/workflows/opencode-pr-review.yml
+++ b/.github/workflows/opencode-pr-review.yml
@@ -363,7 +363,7 @@ jobs:
               issue_number: context.issue.number,
             });
 
-            const aiLabels = ['ai-review-passed', 'ai-review-failed', 'ai-review-warning'];
+            const aiLabels = ['ai-pr-review-passed', 'ai-pr-review-failed', 'ai-pr-review-warning'];
             const toRemove = currentLabels
               .filter(l => aiLabels.includes(l.name))
               .map(l => l.name);
@@ -382,11 +382,11 @@ jobs:
             // Add the appropriate label
             let newLabel;
             if (score >= 80) {
-              newLabel = 'ai-review-passed';
+              newLabel = 'ai-pr-review-passed';
             } else if (score >= 60) {
-              newLabel = 'ai-review-warning';
+              newLabel = 'ai-pr-review-warning';
             } else {
-              newLabel = 'ai-review-failed';
+              newLabel = 'ai-pr-review-failed';
             }
 
             try {

From 3185c31aa329d0cba33cab384522f804fd4992ec Mon Sep 17 00:00:00 2001
From: yhAutomationQA <yhautomationqa@gmail.com>
Date: Sun, 10 May 2026 22:41:24 -0400
Subject: [PATCH 3/9] fix: pass secrets via env blocks instead of run blocks

---
 .github/workflows/opencode-pr-review.yml | 31 ++++++++++++++++--------
 1 file changed, 21 insertions(+), 10 deletions(-)

diff --git a/.github/workflows/opencode-pr-review.yml b/.github/workflows/opencode-pr-review.yml
index 869bf12..c8b8a28 100644
--- a/.github/workflows/opencode-pr-review.yml
+++ b/.github/workflows/opencode-pr-review.yml
@@ -72,31 +72,36 @@ jobs:
     steps:
       # -----------------------------------------------------------------------
       # Step 1: Validate that the AI provider is configured
+      #   Secrets are NOT expanded directly in run blocks (security linter).
+      #   Instead, they are passed via env and checked as env vars.
       # -----------------------------------------------------------------------
       - name: Validate AI provider configuration
         id: validate
+        env:
+          AI_REVIEW_PROVIDER: ${{ secrets.AI_REVIEW_PROVIDER }}
+          OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
+          ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
         run: |
-          PROVIDER="${{ secrets.AI_REVIEW_PROVIDER || 'openai' }}"
-          echo "provider=$PROVIDER" >> $GITHUB_OUTPUT
+          PROVIDER="${AI_REVIEW_PROVIDER:-openai}"
+          echo "provider=$PROVIDER" >> "$GITHUB_OUTPUT"
 
           case "$PROVIDER" in
             openai)
-              if [ -z "${{ secrets.OPENAI_API_KEY }}" ]; then
+              if [ -z "$OPENAI_API_KEY" ]; then
                 echo "Error: OPENAI_API_KEY secret is not set."
                 echo "Set it in your repository secrets: Settings > Secrets and variables > Actions"
                 exit 1
               fi
               ;;
             anthropic)
-              if [ -z "${{ secrets.ANTHROPIC_API_KEY }}" ]; then
+              if [ -z "$ANTHROPIC_API_KEY" ]; then
                 echo "Error: ANTHROPIC_API_KEY secret is not set."
+                echo "Set it in your repository secrets: Settings > Secrets and variables > Actions"
                 exit 1
               fi
               ;;
             opencode)
-              if [ -z "${{ secrets.OPENCODE_API_KEY }}" ]; then
-                echo "Warning: OPENCODE_API_KEY not set — using opencode CLI default."
-              fi
+              echo "Using OpenCode CLI (no API key validation needed)."
               ;;
             *)
               echo "Error: Unknown provider '$PROVIDER'. Supported: openai, anthropic, opencode"
@@ -181,6 +186,10 @@ jobs:
       # -----------------------------------------------------------------------
       - name: Run AI review (OpenCode)
         id: review
+        env:
+          # Secrets passed via env to avoid expanding in run blocks
+          AI_API_KEY: ${{ secrets.OPENAI_API_KEY || secrets.ANTHROPIC_API_KEY || '' }}
+          AI_REVIEW_MODEL: ${{ secrets.AI_REVIEW_MODEL || 'gpt-4o' }}
         run: |
           # Write diff to a temp file so it can be referenced by the prompt
           echo "DIFF_EMPTY=${{ steps.diff.outputs.diff_empty }}" >> $GITHUB_ENV
@@ -215,11 +224,13 @@ jobs:
             fi
           else
             # If opencode CLI is not available, try using an AI API directly
+            # Secrets are passed via environment variables, not on the command line,
+            # to avoid security linter warnings and prevent secret leakage in logs.
             echo "OpenCode CLI not detected. Using AI API fallback..."
-            REVIEW=$(node .github/scripts/ai-review.mjs \
+            REVIEW=$(AI_REVIEW_API_KEY="$AI_API_KEY" \
+              node .github/scripts/ai-review.mjs \
               --provider "${{ steps.validate.outputs.provider }}" \
-              --api-key "${{ secrets.OPENAI_API_KEY || secrets.ANTHROPIC_API_KEY || '' }}" \
-              --model "${{ secrets.AI_REVIEW_MODEL || 'gpt-4o' }}" \
+              --model "$AI_REVIEW_MODEL" \
               --skill-file .ai/skills/playwright-pr-review.md \
               --prompt-file .ai/prompts/pr-review.prompt.md \
               --diff-file /tmp/pr-diff.txt \

From feb6e89699ff5fad93750e1be7507598b54e346c Mon Sep 17 00:00:00 2001
From: yhAutomationQA <yhautomationqa@gmail.com>
Date: Sun, 10 May 2026 22:45:10 -0400
Subject: [PATCH 4/9] =?UTF-8?q?simplify=20to=20OpenAI-only=20=E2=80=94=20r?=
 =?UTF-8?q?emove=20unused=20Anthropic/OpenCode=20branches?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 .github/scripts/ai-review.mjs            | 163 ++++-------------------
 .github/workflows/opencode-pr-review.yml |  82 +++---------
 2 files changed, 44 insertions(+), 201 deletions(-)

diff --git a/.github/scripts/ai-review.mjs b/.github/scripts/ai-review.mjs
index 8e867db..c975e76 100644
--- a/.github/scripts/ai-review.mjs
+++ b/.github/scripts/ai-review.mjs
@@ -1,95 +1,56 @@
 /**
- * ai-review.mjs — AI-powered PR review script
+ * ai-review.mjs — AI-powered PR review script (OpenAI only)
  *
- * This script is the AI provider fallback for the opencode-pr-review workflow.
- * It reads the skill definition, the prompt template, and the git diff,
- * then sends them to an AI API (OpenAI or Anthropic) for analysis.
+ * Reads the skill definition, the prompt template, and the git diff,
+ * then sends them to OpenAI for analysis.
  *
  * Usage:
- *   node .github/scripts/ai-review.mjs \
+ *   OPENAI_API_KEY="sk-..." node .github/scripts/ai-review.mjs \
  *     --provider openai \
- *     --api-key sk-... \
  *     --model gpt-4o \
  *     --skill-file .ai/skills/playwright-pr-review.md \
  *     --prompt-file .ai/prompts/pr-review.prompt.md \
  *     --diff-file /tmp/pr-diff.txt
  *
- * Environment variables (alternative to CLI args):
- *   AI_REVIEW_PROVIDER, AI_REVIEW_API_KEY, AI_REVIEW_MODEL
+ * Environment variables:
+ *   OPENAI_API_KEY  (required)
+ *   AI_REVIEW_MODEL (optional, defaults to gpt-4o)
  */
 
 import fs from 'node:fs';
 import path from 'node:path';
 
-// ---------------------------------------------------------------------------
-// CLI argument parsing
-// ---------------------------------------------------------------------------
 function parseArgs() {
   const args = process.argv.slice(2);
   const opts = {};
-
   for (let i = 0; i < args.length; i++) {
     switch (args[i]) {
-      case '--provider':
-        opts.provider = args[++i];
-        break;
-      case '--api-key':
-        opts.apiKey = args[++i];
-        break;
-      case '--model':
-        opts.model = args[++i];
-        break;
-      case '--skill-file':
-        opts.skillFile = args[++i];
-        break;
-      case '--prompt-file':
-        opts.promptFile = args[++i];
-        break;
-      case '--diff-file':
-        opts.diffFile = args[++i];
-        break;
-      default:
-        // skip
+      case '--provider': opts.provider = args[++i]; break;
+      case '--model':    opts.model = args[++i]; break;
+      case '--skill-file': opts.skillFile = args[++i]; break;
+      case '--prompt-file': opts.promptFile = args[++i]; break;
+      case '--diff-file': opts.diffFile = args[++i]; break;
     }
   }
-
-  // Resolve with environment variable fallbacks
-  opts.provider = opts.provider || process.env.AI_REVIEW_PROVIDER || 'openai';
-  opts.apiKey = opts.apiKey || process.env.AI_REVIEW_API_KEY || '';
-  opts.model = opts.model || process.env.AI_REVIEW_MODEL || 'gpt-4o';
-
+  opts.apiKey = process.env.OPENAI_API_KEY || '';
   return opts;
 }
 
-// ---------------------------------------------------------------------------
-// File reading helpers
-// ---------------------------------------------------------------------------
 function readFileOrExit(filePath, label) {
-  const resolved = path.resolve(filePath);
   try {
-    return fs.readFileSync(resolved, 'utf-8');
+    return fs.readFileSync(path.resolve(filePath), 'utf-8');
   } catch (err) {
-    console.error(`Error reading ${label} at ${resolved}: ${err.message}`);
+    console.error(`Error reading ${label} at ${filePath}: ${err.message}`);
     process.exit(1);
   }
 }
 
-// ---------------------------------------------------------------------------
-// Prompt builder — injects the diff into the prompt template
-// ---------------------------------------------------------------------------
 function buildPrompt(skillContent, promptTemplate, diffContent) {
-  // Prepend the skill rules to the prompt so the AI has full context
   const systemPrompt = `You are a Senior QA Automation Architect. Follow these rules:\n\n${skillContent}`;
-
-  // Inject the diff into the prompt template
   const userPrompt = promptTemplate.replace('{{GIT_DIFF}}', diffContent);
-
   return { systemPrompt, userPrompt };
 }
 
-// ---------------------------------------------------------------------------
-// AI API callers
-// ---------------------------------------------------------------------------
 async function callOpenAI(apiKey, model, systemPrompt, userPrompt) {
   const response = await fetch('https://api.openai.com/v1/chat/completions', {
     method: 'POST',
@@ -98,123 +59,53 @@ async function callOpenAI(apiKey, model, systemPrompt, userPrompt) {
       Authorization: `Bearer ${apiKey}`,
     },
     body: JSON.stringify({
-      model: model,
+      model,
       messages: [
         { role: 'system', content: systemPrompt },
         { role: 'user', content: userPrompt },
       ],
-      temperature: 0.1,       // low temperature for consistent, deterministic review
-      max_tokens: 4096,       // enough for a detailed review
-    }),
-  });
-
-  if (!response.ok) {
-    const errorBody = await response.text();
-    throw new Error(`OpenAI API error ${response.status}: ${errorBody}`);
-  }
-
-  const data = await response.json();
-  return data.choices[0].message.content;
-}
-
-async function callAnthropic(apiKey, model, systemPrompt, userPrompt) {
-  const response = await fetch('https://api.anthropic.com/v1/messages', {
-    method: 'POST',
-    headers: {
-      'Content-Type': 'application/json',
-      'x-api-key': apiKey,
-      'anthropic-version': '2023-06-01',
-    },
-    body: JSON.stringify({
-      model: model,
-      system: systemPrompt,
-      messages: [
-        { role: 'user', content: userPrompt },
-      ],
       temperature: 0.1,
       max_tokens: 4096,
     }),
   });
 
   if (!response.ok) {
-    const errorBody = await response.text();
-    throw new Error(`Anthropic API error ${response.status}: ${errorBody}`);
+    throw new Error(`OpenAI API error ${response.status}: ${await response.text()}`);
   }
 
   const data = await response.json();
-  return data.content[0].text;
+  return data.choices[0].message.content;
 }
 
-// ---------------------------------------------------------------------------
-// Main
-// ---------------------------------------------------------------------------
 async function main() {
   const opts = parseArgs();
 
-  // Validate required inputs
   if (!opts.diffFile || !opts.promptFile || !opts.skillFile) {
-    console.error(
-      'Usage: node ai-review.mjs --provider <provider> --api-key <key> --model <model> ' +
-      '--skill-file <path> --prompt-file <path> --diff-file <path>',
-    );
+    console.error('Missing required arguments: --skill-file, --prompt-file, --diff-file');
     process.exit(1);
   }
 
-  // Verify the diff actually has content
   const diffContent = readFileOrExit(opts.diffFile, 'diff file');
   if (!diffContent.trim() || diffContent.trim() === '(no diff — no Playwright/TypeScript files changed)') {
-    console.log('## AI PR Review');
-    console.log('');
-    console.log('**No issues found.** No Playwright/TypeScript code changes detected in this diff.');
-    console.log('');
-    console.log('## PR Quality Score: 100/100 — Excellent');
-    console.log('');
-    console.log('| Severity | Count |');
-    console.log('|----------|-------|');
-    console.log('| Critical | 0 |');
-    console.log('| Major | 0 |');
-    console.log('| Minor | 0 |');
-    console.log('| Suggestion | 0 |');
+    console.log('## AI PR Review\n\n**No issues found.** No Playwright/TypeScript code changes detected.\n\n## PR Quality Score: 100/100 — Excellent\n\n| Severity | Count |\n|----------|-------|\n| Critical | 0 |\n| Major | 0 |\n| Minor | 0 |\n| Suggestion | 0 |');
     return;
   }
 
-  const skillContent = readFileOrExit(opts.skillFile, 'skill file');
-  const promptTemplate = readFileOrExit(opts.promptFile, 'prompt template');
-
-  const { systemPrompt, userPrompt } = buildPrompt(skillContent, promptTemplate, diffContent);
-
   if (!opts.apiKey) {
-    console.error('Error: No API key provided. Set AI_REVIEW_API_KEY environment variable or pass --api-key.');
+    console.error('Error: OPENAI_API_KEY environment variable is not set.');
     process.exit(1);
   }
 
-  // Call the appropriate AI provider
-  let reviewText;
-  switch (opts.provider) {
-    case 'openai':
-      reviewText = await callOpenAI(opts.apiKey, opts.model, systemPrompt, userPrompt);
-      break;
-    case 'anthropic':
-      reviewText = await callAnthropic(opts.apiKey, opts.model, systemPrompt, userPrompt);
-      break;
-    default:
-      console.error(`Unsupported provider: ${opts.provider}. Use 'openai' or 'anthropic'.`);
-      process.exit(1);
-  }
+  const skillContent = readFileOrExit(opts.skillFile, 'skill file');
+  const promptTemplate = readFileOrExit(opts.promptFile, 'prompt template');
+  const { systemPrompt, userPrompt } = buildPrompt(skillContent, promptTemplate, diffContent);
 
-  // Print the review to stdout (captured by the workflow)
+  const reviewText = await callOpenAI(opts.apiKey, opts.model || 'gpt-4o', systemPrompt, userPrompt);
   console.log(reviewText);
 }
 
 main().catch((err) => {
-  console.error('AI review script failed:', err.message);
-  // Output a fallback review so the workflow doesn't break
-  console.log('## AI PR Review');
-  console.log('');
-  console.log('**Review could not be completed.**');
-  console.log('');
-  console.log(`Error: ${err.message}`);
-  console.log('');
-  console.log('## PR Quality Score: N/A');
+  console.error('AI review failed:', err.message);
+  console.log('## AI PR Review\n\n**Review could not be completed.**\n\nError: ' + err.message);
   process.exit(1);
 });
diff --git a/.github/workflows/opencode-pr-review.yml b/.github/workflows/opencode-pr-review.yml
index c8b8a28..67d0b1c 100644
--- a/.github/workflows/opencode-pr-review.yml
+++ b/.github/workflows/opencode-pr-review.yml
@@ -75,39 +75,17 @@ jobs:
       #   Secrets are NOT expanded directly in run blocks (security linter).
       #   Instead, they are passed via env and checked as env vars.
       # -----------------------------------------------------------------------
-      - name: Validate AI provider configuration
+      - name: Validate OpenAI API key
         id: validate
         env:
-          AI_REVIEW_PROVIDER: ${{ secrets.AI_REVIEW_PROVIDER }}
           OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
-          ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
         run: |
-          PROVIDER="${AI_REVIEW_PROVIDER:-openai}"
-          echo "provider=$PROVIDER" >> "$GITHUB_OUTPUT"
-
-          case "$PROVIDER" in
-            openai)
-              if [ -z "$OPENAI_API_KEY" ]; then
-                echo "Error: OPENAI_API_KEY secret is not set."
-                echo "Set it in your repository secrets: Settings > Secrets and variables > Actions"
-                exit 1
-              fi
-              ;;
-            anthropic)
-              if [ -z "$ANTHROPIC_API_KEY" ]; then
-                echo "Error: ANTHROPIC_API_KEY secret is not set."
-                echo "Set it in your repository secrets: Settings > Secrets and variables > Actions"
-                exit 1
-              fi
-              ;;
-            opencode)
-              echo "Using OpenCode CLI (no API key validation needed)."
-              ;;
-            *)
-              echo "Error: Unknown provider '$PROVIDER'. Supported: openai, anthropic, opencode"
-              exit 1
-              ;;
-          esac
+          if [ -z "$OPENAI_API_KEY" ]; then
+            echo "Error: OPENAI_API_KEY secret is not set."
+            echo "Set it in your repository secrets: Settings > Secrets and variables > Actions"
+            exit 1
+          fi
+          echo "provider=openai" >> "$GITHUB_OUTPUT"
 
       # -----------------------------------------------------------------------
       # Step 2: Check out the PR branch with full git history
@@ -187,11 +165,9 @@ jobs:
       - name: Run AI review (OpenCode)
         id: review
         env:
-          # Secrets passed via env to avoid expanding in run blocks
-          AI_API_KEY: ${{ secrets.OPENAI_API_KEY || secrets.ANTHROPIC_API_KEY || '' }}
+          OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
           AI_REVIEW_MODEL: ${{ secrets.AI_REVIEW_MODEL || 'gpt-4o' }}
         run: |
-          # Write diff to a temp file so it can be referenced by the prompt
           echo "DIFF_EMPTY=${{ steps.diff.outputs.diff_empty }}" >> $GITHUB_ENV
 
           if [ "${{ steps.diff.outputs.diff_empty }}" = "true" ]; then
@@ -201,41 +177,17 @@ jobs:
             exit 0
           fi
 
-          # Write the diff to a temporary file for the review script
           echo "${{ steps.diff.outputs.diff_content }}" > /tmp/pr-diff.txt
 
-          # Check for opencode CLI
-          if command -v npx &> /dev/null && npx --yes opencode --help &> /dev/null 2>&1; then
-            # Run the review using opencode with the skill and prompt
-            # This approach loads the skill, reads the prompt template, injects the diff
-            echo "Running review via OpenCode..."
-            npx opencode review \
-              --skill .ai/skills/playwright-pr-review.md \
-              --prompt .ai/prompts/pr-review.prompt.md \
-              --diff /tmp/pr-diff.txt \
-              --output /tmp/review-output.txt \
-              --format markdown \
-              2>&1 || true
-
-            if [ -f /tmp/review-output.txt ]; then
-              REVIEW=$(cat /tmp/review-output.txt)
-            else
-              REVIEW="(OpenCode review produced no output)"  
-            fi
-          else
-            # If opencode CLI is not available, try using an AI API directly
-            # Secrets are passed via environment variables, not on the command line,
-            # to avoid security linter warnings and prevent secret leakage in logs.
-            echo "OpenCode CLI not detected. Using AI API fallback..."
-            REVIEW=$(AI_REVIEW_API_KEY="$AI_API_KEY" \
-              node .github/scripts/ai-review.mjs \
-              --provider "${{ steps.validate.outputs.provider }}" \
-              --model "$AI_REVIEW_MODEL" \
-              --skill-file .ai/skills/playwright-pr-review.md \
-              --prompt-file .ai/prompts/pr-review.prompt.md \
-              --diff-file /tmp/pr-diff.txt \
-              2>&1 || echo "(AI review script failed — see logs)")
-          fi
+          echo "Running AI review via OpenAI..."
+          REVIEW=$(OPENAI_API_KEY="$OPENAI_API_KEY" \
+            node .github/scripts/ai-review.mjs \
+            --provider openai \
+            --model "$AI_REVIEW_MODEL" \
+            --skill-file .ai/skills/playwright-pr-review.md \
+            --prompt-file .ai/prompts/pr-review.prompt.md \
+            --diff-file /tmp/pr-diff.txt \
+            2>&1 || echo "(AI review script failed — see logs)")
 
           # Write review output
           echo "review_output<<EOF_REVIEW" >> $GITHUB_OUTPUT

From a2d87d3cf5fd871c589663105dfc7f91e2186a7d Mon Sep 17 00:00:00 2001
From: yhAutomationQA <yhautomationqa@gmail.com>
Date: Sun, 10 May 2026 22:48:23 -0400
Subject: [PATCH 5/9] fix: write diff to file via base64 to avoid bash special
 char errors

---
 .github/workflows/opencode-pr-review.yml | 19 +++++++++----------
 1 file changed, 9 insertions(+), 10 deletions(-)

diff --git a/.github/workflows/opencode-pr-review.yml b/.github/workflows/opencode-pr-review.yml
index 67d0b1c..6aa25f6 100644
--- a/.github/workflows/opencode-pr-review.yml
+++ b/.github/workflows/opencode-pr-review.yml
@@ -125,21 +125,21 @@ jobs:
             ':!*.ttf' \
             2>/dev/null || echo "(empty diff or error)")
 
-          # Handle empty or large diffs
+          # Write diff to file to avoid bash special character issues when inlining
+          # Base64 encoding ensures parentheses, quotes, backticks, etc. don't break the shell
           if [ -z "$DIFF" ]; then
             echo "diff_empty=true" >> $GITHUB_OUTPUT
-            echo "diff_content=(no diff — no Playwright/TypeScript files changed)" >> $GITHUB_OUTPUT
+            echo -n "" | base64 > /tmp/pr-diff.b64
           else
             # Truncate diff to prevent token overflow (conservative limit)
             MAX_DIFF_LENGTH=$(( 4 * 1024 * 50 ))  # ~50K tokens worth of diff
             if [ ${#DIFF} -gt $MAX_DIFF_LENGTH ]; then
               DIFF="${DIFF:0:$MAX_DIFF_LENGTH}"
-              DIFF+="\n\n[... diff truncated at ${MAX_DIFF_LENGTH} characters — only the first part was analyzed ...]"
+              DIFF+=$'\n\n[... diff truncated at '"${MAX_DIFF_LENGTH}"$' characters — only the first part was analyzed ...]'
             fi
+            # Write base64-encoded diff to file (safe from shell interpretation)
+            echo -n "$DIFF" | base64 > /tmp/pr-diff.b64
             echo "diff_empty=false" >> $GITHUB_OUTPUT
-            echo "diff_content<<EOF_DIFF" >> $GITHUB_OUTPUT
-            echo "$DIFF" >> $GITHUB_OUTPUT
-            echo "EOF_DIFF" >> $GITHUB_OUTPUT
           fi
 
       # -----------------------------------------------------------------------
@@ -168,17 +168,16 @@ jobs:
           OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
           AI_REVIEW_MODEL: ${{ secrets.AI_REVIEW_MODEL || 'gpt-4o' }}
         run: |
-          echo "DIFF_EMPTY=${{ steps.diff.outputs.diff_empty }}" >> $GITHUB_ENV
+          # Decode the diff from base64 (avoids bash special character issues)
+          base64 -d /tmp/pr-diff.b64 > /tmp/pr-diff.txt 2>/dev/null || true
 
-          if [ "${{ steps.diff.outputs.diff_empty }}" = "true" ]; then
+          if [ ! -s /tmp/pr-diff.txt ]; then
             echo "review_output=No issues found. No Playwright/TypeScript code changed." >> $GITHUB_OUTPUT
             echo "review_score=100" >> $GITHUB_OUTPUT
             echo "review_rating=Excellent" >> $GITHUB_OUTPUT
             exit 0
           fi
 
-          echo "${{ steps.diff.outputs.diff_content }}" > /tmp/pr-diff.txt
-
           echo "Running AI review via OpenAI..."
           REVIEW=$(OPENAI_API_KEY="$OPENAI_API_KEY" \
             node .github/scripts/ai-review.mjs \

From eaf14f518081bbcd137cb66aa0089a22e8b72618 Mon Sep 17 00:00:00 2001
From: yhAutomationQA <yhautomationqa@gmail.com>
Date: Sun, 10 May 2026 22:51:56 -0400
Subject: [PATCH 6/9] fix: write review output to files to avoid JS template
 literal breakage from backticks in markdown

---
 .github/workflows/opencode-pr-review.yml | 69 ++++++++----------------
 1 file changed, 23 insertions(+), 46 deletions(-)

diff --git a/.github/workflows/opencode-pr-review.yml b/.github/workflows/opencode-pr-review.yml
index 6aa25f6..110d83f 100644
--- a/.github/workflows/opencode-pr-review.yml
+++ b/.github/workflows/opencode-pr-review.yml
@@ -172,9 +172,9 @@ jobs:
           base64 -d /tmp/pr-diff.b64 > /tmp/pr-diff.txt 2>/dev/null || true
 
           if [ ! -s /tmp/pr-diff.txt ]; then
-            echo "review_output=No issues found. No Playwright/TypeScript code changed." >> $GITHUB_OUTPUT
-            echo "review_score=100" >> $GITHUB_OUTPUT
-            echo "review_rating=Excellent" >> $GITHUB_OUTPUT
+            echo -n "No issues found. No Playwright/TypeScript code changed." > /tmp/review-output.md
+            echo -n "100" > /tmp/review-score.txt
+            echo -n "Excellent" > /tmp/review-rating.txt
             exit 0
           fi
 
@@ -186,32 +186,28 @@ jobs:
             --skill-file .ai/skills/playwright-pr-review.md \
             --prompt-file .ai/prompts/pr-review.prompt.md \
             --diff-file /tmp/pr-diff.txt \
-            2>&1 || echo "(AI review script failed — see logs)")
+            2>&1 || echo "(AI review script failed -- see logs)")
 
-          # Write review output
-          echo "review_output<<EOF_REVIEW" >> $GITHUB_OUTPUT
-          echo "$REVIEW" >> $GITHUB_OUTPUT
-          echo "EOF_REVIEW" >> $GITHUB_OUTPUT
+          # Write review output to file (avoids shell/JS special character issues from $GITHUB_OUTPUT)
+          echo "$REVIEW" > /tmp/review-output.md
 
           # Extract score from review output (format: "## PR Quality Score: <N>/100")
           SCORE=$(echo "$REVIEW" | grep -oP 'PR Quality Score:\s*\K(\d+)' || echo "")
           if [ -z "$SCORE" ]; then
-            # Fallback: parse severity counts to compute score
             CRIT=$(echo "$REVIEW" | grep -c '^|.*Critical.*|' || echo 0)
             MAJ=$(echo "$REVIEW" | grep -c '^|.*Major.*|' || echo 0)
             MIN=$(echo "$REVIEW" | grep -c '^|.*Minor.*|' || echo 0)
             SCORE=$(( 100 - (CRIT * 15) - (MAJ * 10) - (MIN * 5) ))
             [ "$SCORE" -lt 0 ] && SCORE=0
           fi
-          echo "review_score=$SCORE" >> $GITHUB_OUTPUT
+          echo -n "$SCORE" > /tmp/review-score.txt
 
-          # Compute rating
           if [ "$SCORE" -eq 100 ]; then RATING="Excellent"
           elif [ "$SCORE" -ge 80 ]; then RATING="Good"
           elif [ "$SCORE" -ge 60 ]; then RATING="Needs Work"
           else RATING="Failing"
           fi
-          echo "review_rating=$RATING" >> $GITHUB_OUTPUT
+          echo -n "$RATING" > /tmp/review-rating.txt
 
       # -----------------------------------------------------------------------
       # Step 6: Post review as a PR comment
@@ -222,13 +218,19 @@ jobs:
         id: comment
         with:
           script: |
+            const fs = require('fs');
+
+            // Read review artifacts from files (avoids shell/JS expansion issues)
+            const reviewOutput = fs.readFileSync('/tmp/review-output.md', 'utf8');
+            const score = fs.readFileSync('/tmp/review-score.txt', 'utf8');
+            const rating = fs.readFileSync('/tmp/review-rating.txt', 'utf8');
+
             const { data: comments } = await github.rest.issues.listComments({
               owner: context.repo.owner,
               repo: context.repo.repo,
               issue_number: context.issue.number,
             });
 
-            // Find and delete any previous AI review comment
             const marker = '<!-- opencode-pr-review -->';
             for (const comment of comments) {
               if (comment.body && comment.body.includes(marker)) {
@@ -240,22 +242,16 @@ jobs:
               }
             }
 
-            // Build the review body
-            const reviewOutput = `${{ steps.review.outputs.review_output }}`;
-            const score = `${{ steps.review.outputs.review_score }}`;
-            const rating = `${{ steps.review.outputs.review_rating }}`;
-
-            const header = `## 🤖 AI PR Review\n\n`;
+            const header = '## 🤖 AI PR Review\n\n';
             const footer = `\n\n---\n*This review was generated automatically by the AI PR Review workflow. Human approval is required before merge.*\n${marker}`;
 
             let body;
             if (reviewOutput.includes('No issues found')) {
-              body = header + `**No issues found.** PR adheres to framework standards.\n\n**PR Quality Score:** ${score}/100 — ${rating}` + footer;
+              body = header + '**No issues found.** PR adheres to framework standards.\n\n**PR Quality Score:** ' + score + '/100 — ' + rating + footer;
             } else {
               body = header + reviewOutput + footer;
             }
 
-            // Post the new review comment
             await github.rest.issues.createComment({
               owner: context.repo.owner,
               repo: context.repo.repo,
@@ -274,8 +270,9 @@ jobs:
         uses: actions/github-script@v7
         with:
           script: |
-            const score = parseInt('${{ steps.review.outputs.review_score }}' || '100', 10);
-            const rating = '${{ steps.review.outputs.review_rating }}' || 'Good';
+            const fs = require('fs');
+            const score = parseInt(fs.readFileSync('/tmp/review-score.txt', 'utf8') || '100', 10);
+            const rating = fs.readFileSync('/tmp/review-rating.txt', 'utf8') || 'Good';
             const headSha = '${{ github.event.pull_request.head.sha }}';
 
             // Determine conclusion based on score
@@ -315,7 +312,8 @@ jobs:
         uses: actions/github-script@v7
         with:
           script: |
-            const score = parseInt('${{ steps.review.outputs.review_score }}' || '100', 10);
+            const fs = require('fs');
+            const score = parseInt(fs.readFileSync('/tmp/review-score.txt', 'utf8') || '100', 10);
             const labels = new Set();
 
             // Remove existing AI review labels
@@ -363,25 +361,4 @@ jobs:
               console.log('Could not add label. Create it in your repository settings.');
             }
 
-  # ===========================================================================
-  # Job 2: summary-report (optional)
-  #   Posts a summary to the workflow run for easy visibility.
-  # ===========================================================================
-  summary-report:
-    needs: review
-    if: always()
-    runs-on: ubuntu-latest
-    steps:
-      - name: Post workflow summary
-        run: |
-          echo "## AI PR Review Summary" >> $GITHUB_STEP_SUMMARY
-          echo "" >> $GITHUB_STEP_SUMMARY
-          echo "| Metric | Value |" >> $GITHUB_STEP_SUMMARY
-          echo "|--------|-------|" >> $GITHUB_STEP_SUMMARY
-          echo "| PR | #${{ github.event.pull_request.number }} |" >> $GITHUB_STEP_SUMMARY
-          echo "| Branch | ${{ github.event.pull_request.head.ref }} → ${{ github.event.pull_request.base.ref }} |" >> $GITHUB_STEP_SUMMARY
-          echo "| Author | ${{ github.event.pull_request.user.login }} |" >> $GITHUB_STEP_SUMMARY
-          echo "| Quality Score | ${{ needs.review.outputs.review_score || 'N/A' }} / 100 |" >> $GITHUB_STEP_SUMMARY
-          echo "| Rating | ${{ needs.review.outputs.review_rating || 'N/A' }} |" >> $GITHUB_STEP_SUMMARY
-          echo "" >> $GITHUB_STEP_SUMMARY
-          echo "> **Note:** This is an automated review. A human maintainer must approve the PR before merge." >> $GITHUB_STEP_SUMMARY
+

From 2f574a071fa7504a03b8bff47228bd7c177a7866 Mon Sep 17 00:00:00 2001
From: yhAutomationQA <yhautomationqa@gmail.com>
Date: Sun, 10 May 2026 22:55:39 -0400
Subject: [PATCH 7/9] rename workflow file: opencode -> ai

---
 .github/workflows/{opencode-pr-review.yml => ai-pr-review.yml} | 0
 1 file changed, 0 insertions(+), 0 deletions(-)
 rename .github/workflows/{opencode-pr-review.yml => ai-pr-review.yml} (100%)

diff --git a/.github/workflows/opencode-pr-review.yml b/.github/workflows/ai-pr-review.yml
similarity index 100%
rename from .github/workflows/opencode-pr-review.yml
rename to .github/workflows/ai-pr-review.yml

From 1ea2e8a879bba580e333864d17237e535b8161d1 Mon Sep 17 00:00:00 2001
From: yhAutomationQA <yhautomationqa@gmail.com>
Date: Sun, 10 May 2026 22:56:28 -0400
Subject: [PATCH 8/9] remove remaining opencode references

---
 .github/workflows/ai-pr-review.yml | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/.github/workflows/ai-pr-review.yml b/.github/workflows/ai-pr-review.yml
index 110d83f..72d8172 100644
--- a/.github/workflows/ai-pr-review.yml
+++ b/.github/workflows/ai-pr-review.yml
@@ -1,5 +1,5 @@
 # =============================================================================
-# opencode-pr-review.yml — AI-Powered PR Review for Playwright + TypeScript
+# ai-pr-review.yml — AI-Powered PR Review for Playwright + TypeScript
 # =============================================================================
 #
 # This workflow triggers on pull requests to `main`, analyzes the git diff
@@ -231,7 +231,7 @@ jobs:
               issue_number: context.issue.number,
             });
 
-            const marker = '<!-- opencode-pr-review -->';
+            const marker = '<!-- ai-pr-review -->';
             for (const comment of comments) {
               if (comment.body && comment.body.includes(marker)) {
                 await github.rest.issues.deleteComment({

From 4d0e4d27edba776ae17b90c0a458d9bbd4a87680 Mon Sep 17 00:00:00 2001
From: yhAutomationQA <yhautomationqa@gmail.com>
Date: Sun, 10 May 2026 23:05:11 -0400
Subject: [PATCH 9/9] remove remaining opencode references from workflow
 comments

---
 .github/workflows/ai-pr-review.yml | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/.github/workflows/ai-pr-review.yml b/.github/workflows/ai-pr-review.yml
index 72d8172..3b468e5 100644
--- a/.github/workflows/ai-pr-review.yml
+++ b/.github/workflows/ai-pr-review.yml
@@ -3,7 +3,7 @@
 # =============================================================================
 #
 # This workflow triggers on pull requests to `main`, analyzes the git diff
-# using an AI agent (OpenCode or OpenAI-compatible API), and posts a
+# using OpenAI's API, and posts a
 # structured code review with severity-labeled findings and a quality score.
 #
 # Design Principles:
@@ -158,11 +158,10 @@ jobs:
           PLAYWRIGHT_SKIP_BROWSER_DOWNLOAD: 1
 
       # -----------------------------------------------------------------------
-      # Step 5 (Option A): Run AI review via OpenCode CLI
-      #     Uses the OpenCode CLI tool to run the PR review skill against the
+      # Step 5: Run AI review via OpenAI API
       #     prompt template. Falls back to API-based review if not configured.
       # -----------------------------------------------------------------------
-      - name: Run AI review (OpenCode)
+      - name: Run AI review
         id: review
         env:
           OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}