PyBreeze 1.0.4__py3-none-any.whl

automation_ide/automation_editor_ui/extend_ai_gui/prompt_edit_gui/cot_code_review_prompt_templates/judge.py

@@ -0,0 +1,91 @@
+JUDGE_TEMPLATE = """
+# Code Review Comment Evaluation Template (Enhanced)
+
+Please evaluate the review comments focusing on how well they address important issues in the code, especially leveraging code smell and linter messages.  
+The input may contain multiple independent review reports.
+
+Score range: 1–100
+
+Five evaluation dimensions:
+### 1. Readability
+    - 1–20: Comments are very hard to understand, poorly structured, confusing language.
+    - 21–40: Some parts are readable, but many unclear sections remain.
+    - 41–60: Comments are generally clear, but structure or phrasing needs improvement.
+    - 61–80: Comments are well-structured, consistent, and easy to follow.
+    - 81–100: Comments are highly readable, elegantly phrased, and well-organized.
+
+### 2. Constructiveness (Maintainability)
+    - 1–20: Comments lack constructive suggestions, no improvement direction.
+    - 21–40: Comments provide partial suggestions, but vague or impractical.
+    - 41–60: Comments offer basic improvement ideas, somewhat helpful.
+    - 61–80: Comments are specific and actionable, clearly guiding improvements.
+    - 81–100: Comments are highly constructive, offering clear and practical improvement paths.
+
+### 3. Correctness
+    - 1–20: Comments contain errors or misleading advice.
+    - 21–40: Mostly correct, but important issues are overlooked.
+    - 41–60: Largely correct, with only minor gaps.
+    - 61–80: Correct and reasonable, with small room for refinement.
+    - 81–100: Fully correct, logically sound, and precise in identifying issues.
+
+### 4. Multi-Review Coverage, Structural Independence & Extractability
+
+    > Evaluate how well the comments cover important issues across multiple review reports, and whether each comment block is structurally independent, self-contained, and understandable on its own.
+
+    #### Scoring Criteria
+
+    - **1–20**  
+      Rarely addresses important issues; structure is disorganized; comments heavily depend on surrounding context; blocks cannot be read independently; unclear linkage to specific code smells or linter messages.
+
+    - **21–40**  
+      Addresses some issues but misses many key points; unclear boundaries between sections; frequent cross-references required for understanding; difficult to isolate specific issue blocks.
+
+    - **41–60**  
+      Covers many important issues; basic structure is present; some blocks can be read independently, but certain sections still rely on context or lack completeness.
+
+    - **61–80**  
+      Most key issues are addressed; comment blocks are clearly structured with defined themes; most sections can be independently read and understood (problem + reasoning + suggestion); code smell / linter-related blocks are reasonably extractable.
+
+    - **81–100**  
+      Thoroughly addresses key issues; each comment block is **independent, complete, and self-contained** (including problem description, impact explanation, and actionable improvement suggestions);  
+      clearly segmented structure; any block can be extracted without losing meaning;  
+      strongly aligned with specific code smells or linter messages; highly readable and maintainable.
+    ---
+    ### Additional Evaluation Criteria
+
+    When scoring this dimension, explicitly check whether each comment block:
+    - Has a clear title or thematic focus  
+    - Clearly identifies the issue source (e.g., specific code smell or linter message)  
+    - Explains impact or risk  
+    - Provides concrete and actionable improvement suggestions  
+    - Does not rely on other sections for comprehension  
+
+    Also assess whether:
+    - A single comment block can be extracted and shared independently without losing clarity  
+    - Code smell or linter-related sections can be isolated for structured analysis  
+    - Summary sections and detailed comments are clearly distinguished  
+
+### 5. Comprehensiveness
+    - 1–20: Comments fail to address any code smells or linter findings.
+    - 21–40: Comments mention at least one code smell or linter warning.
+    - 41–60: Comments cover some code smells or linter findings.
+    - 61–80: Comments cover most code smells and linter findings.
+    - 81–100: Comments comprehensively address all code smells and linter findings, with improvement suggestions.
+
+Reviewers should:
+- Assign a score (1–100) for each dimension.
+- Provide brief reasoning for each score.
+- Conclude with an average score and overall recommendation.
+
+## Review Comment:
+{review_comment}
+
+## Code Smells:
+{code_smell_detector_messages}
+
+## Linter Messages:
+{linter_messages}
+
+## Origin code
+{code_diff}
+"""

automation_ide/automation_editor_ui/extend_ai_gui/prompt_edit_gui/cot_code_review_prompt_templates/judge_single_review.py

@@ -0,0 +1,85 @@
+JUDGE_SINGLE_REVIEW_TEMPLATE = """
+# Code Review Comment Evaluation Template (Enhanced)
+
+Please evaluate the review comments focusing on how well they address important issues in the code, especially leveraging code smell and linter messages.  
+The input may contain multiple independent review reports.
+
+Score range: 1–100
+
+Five evaluation dimensions:
+### 1. Readability
+    - 1–20: Comments are very hard to understand, poorly structured, confusing language.
+    - 21–40: Some parts are readable, but many unclear sections remain.
+    - 41–60: Comments are generally clear, but structure or phrasing needs improvement.
+    - 61–80: Comments are well-structured, consistent, and easy to follow.
+    - 81–100: Comments are highly readable, elegantly phrased, and well-organized.
+
+### 2. Constructiveness (Maintainability)
+    - 1–20: Comments lack constructive suggestions, no improvement direction.
+    - 21–40: Comments provide partial suggestions, but vague or impractical.
+    - 41–60: Comments offer basic improvement ideas, somewhat helpful.
+    - 61–80: Comments are specific and actionable, clearly guiding improvements.
+    - 81–100: Comments are highly constructive, offering clear and practical improvement paths.
+
+### 3. Correctness
+    - 1–20: Comments contain errors or misleading advice.
+    - 21–40: Mostly correct, but important issues are overlooked.
+    - 41–60: Largely correct, with only minor gaps.
+    - 61–80: Correct and reasonable, with small room for refinement.
+    - 81–100: Fully correct, logically sound, and precise in identifying issues.
+
+### 4. Multi-Review Coverage, Structural Independence & Extractability
+
+    > Evaluate how well the comments cover important issues across multiple review reports, and whether each comment block is structurally independent, self-contained, and understandable on its own.
+
+    #### Scoring Criteria
+
+    - **1–20**  
+      Rarely addresses important issues; structure is disorganized; comments heavily depend on surrounding context; blocks cannot be read independently; unclear linkage to specific code smells or linter messages.
+
+    - **21–40**  
+      Addresses some issues but misses many key points; unclear boundaries between sections; frequent cross-references required for understanding; difficult to isolate specific issue blocks.
+
+    - **41–60**  
+      Covers many important issues; basic structure is present; some blocks can be read independently, but certain sections still rely on context or lack completeness.
+
+    - **61–80**  
+      Most key issues are addressed; comment blocks are clearly structured with defined themes; most sections can be independently read and understood (problem + reasoning + suggestion); code smell / linter-related blocks are reasonably extractable.
+
+    - **81–100**  
+      Thoroughly addresses key issues; each comment block is **independent, complete, and self-contained** (including problem description, impact explanation, and actionable improvement suggestions);  
+      clearly segmented structure; any block can be extracted without losing meaning;  
+      strongly aligned with specific code smells or linter messages; highly readable and maintainable.
+    ---
+    ### Additional Evaluation Criteria
+
+    When scoring this dimension, explicitly check whether each comment block:
+    - Has a clear title or thematic focus  
+    - Clearly identifies the issue source (e.g., specific code smell or linter message)  
+    - Explains impact or risk  
+    - Provides concrete and actionable improvement suggestions  
+    - Does not rely on other sections for comprehension  
+
+    Also assess whether:
+    - A single comment block can be extracted and shared independently without losing clarity  
+    - Code smell or linter-related sections can be isolated for structured analysis  
+    - Summary sections and detailed comments are clearly distinguished  
+
+### 5. Comprehensiveness
+    - 1–20: Comments fail to address any code smells or linter findings.
+    - 21–40: Comments mention at least one code smell or linter warning.
+    - 41–60: Comments cover some code smells or linter findings.
+    - 61–80: Comments cover most code smells and linter findings.
+    - 81–100: Comments comprehensively address all code smells and linter findings, with improvement suggestions.
+
+Reviewers should:
+- Assign a score (1–100) for each dimension.
+- Provide brief reasoning for each score.
+- Conclude with an average score and overall recommendation.
+
+## Review Comment:
+{review_comment}
+
+## Origin code
+{code_diff}
+"""

automation_ide/automation_editor_ui/extend_ai_gui/prompt_edit_gui/cot_code_review_prompt_templates/linter.py

@@ -0,0 +1,17 @@
+LINTER_TEMPLATE = """
+    You are a strict code linter. 
+    Your task is to analyze the given source code and produce structured linter_messages. 
+    Follow these rules:
+
+    1. Do not rewrite or fix the code — only report issues.
+    2. Each linter_message must include:
+       - rule_id: A short identifier for the rule violated (e.g., "no-unused-vars").
+       - severity: One of ["error", "warning", "info"].
+       - message: A clear explanation of the issue.
+       - line: The line number where the issue occurs.
+       - suggestion: A concise recommendation for improvement.
+    3. If no issues are found, return an empty list.
+
+    Now analyze the following code:
+    {code_diff}
+"""

automation_ide/automation_editor_ui/extend_ai_gui/prompt_edit_gui/cot_code_review_prompt_templates/step_by_step_analysis.py

@@ -0,0 +1,35 @@
+STEP_BY_STEP_ANALYSIS_TEMPLATE = """
+You are a code quality reviewer. 
+Analyze the code smell, and linter message step by step.
+
+Instructions:
+1. **Identify the Issue**  
+   - Restate the code smell or lint message in plain English.  
+   - Explain what it means in the context of software engineering.
+
+2. **Root Cause Analysis**  
+   - Describe why this issue occurs.  
+   - Point out the underlying coding practice or design flaw.
+
+3. **Impact Assessment**  
+   - Explain the potential risks (e.g., maintainability, readability, performance, security).  
+   - Clarify how severe the issue is.
+
+4. **Suggested Fix**  
+   - Provide a concise, actionable recommendation.  
+   - If relevant, show a short code snippet with the corrected approach.
+
+5. **Best Practice Note**  
+   - Mention a general guideline or principle (e.g., SOLID, DRY, naming conventions) that helps prevent similar issues.
+
+## Linter Result
+{linter_result}
+
+## Code Smell Result
+{code_smell_result}
+
+Output Format:
+- Use numbered steps for each lint message or code smell.  
+- Keep explanations simple but professional.  
+- Provide examples where helpful.   
+"""

automation_ide/automation_editor_ui/extend_ai_gui/prompt_edit_gui/cot_code_review_prompt_templates/total_summary.py

@@ -0,0 +1,66 @@
+TOTAL_SUMMARY_TEMPLATE = """
+# PR Total Summary Instructions
+
+You are a senior code reviewer generating the **final PR Total Summary**.
+You MUST base your analysis **strictly and exclusively** on the information provided below.
+Do NOT introduce assumptions, external standards, or unstated context.
+
+Your task is to synthesize all provided review artifacts into a single, well-reasoned conclusion.
+
+## Requirements
+
+- Use **all provided inputs** as evidence:
+  - First Code Review Result
+  - First Summary Result
+  - Linter Result
+  - Code Smell Result
+  - Code Diff
+- The final summary must clearly reflect **actual findings** from the above inputs.
+- Avoid generic statements such as “code quality is good” unless directly supported by the data.
+- Do NOT repeat earlier comments verbatim; instead, **aggregate and interpret** them.
+- If there are conflicting signals (e.g., linter passes but code smells exist), explicitly reconcile them.
+- If critical issues exist, they must directly impact the final recommendation.
+
+## Output Format
+
+Respond in **structured bullet points**, professional and concise, covering only:
+
+1. **Overall conclusion**
+   - Does the PR meet merge criteria based on the evidence?
+   - Mention any blocking vs non-blocking concerns.
+
+2. **Comprehensive evaluation**
+   - Code quality and correctness (derived from review + diff)
+   - Maintainability and design concerns (derived from code smells / structure)
+   - Consistency with existing patterns or standards (only if inferable from the diff)
+
+3. **Final decision recommendation**
+   - One of: Approve merge / Request changes / Comment only
+   - Justify the decision using concrete findings.
+
+4. **Team follow-up (if applicable)**
+   - Specific, actionable next steps grounded in the findings
+   - Omit this section if no follow-up is needed.
+
+Constraints:
+- Do NOT add new issues not present in the provided inputs.
+- Do NOT speculate about runtime behavior or architecture beyond what the diff shows.
+- Focus solely on final judgment, not detailed line-by-line critique.
+
+---
+
+## First Code Review Result
+{first_code_review}
+
+## First Summary Result
+{first_summary}
+
+## Linter Result
+{linter_result}
+
+## Code Smell Result
+{code_smell_result}
+
+## Code Diff
+{code_diff}
+"""

automation_ide/automation_editor_ui/extend_ai_gui/prompt_edit_gui/cot_prompt_editor_widget.py

@@ -0,0 +1,136 @@
+import os
+
+from PySide6.QtCore import QFileSystemWatcher
+from PySide6.QtWidgets import (
+    QWidget, QVBoxLayout, QHBoxLayout,
+    QComboBox, QTextEdit, QPushButton, QGroupBox, QMessageBox
+)
+from je_editor import language_wrapper
+
+from automation_ide.automation_editor_ui.extend_ai_gui.ai_gui_global_variable import COT_TEMPLATE_FILES, \
+    COT_TEMPLATE_RELATION
+
+
+class CoTPromptEditor(QWidget):
+    def __init__(self, prompt_files=None, parent=None):
+        super().__init__(parent)
+        self.prompt_files = prompt_files or COT_TEMPLATE_FILES
+
+        # 對應檔案名稱與模板內容
+        self.templates = COT_TEMPLATE_RELATION
+
+        self.setWindowTitle(language_wrapper.language_word_dict.get(
+            "cot_cot_prompt_editor_window_title"
+        ))  # 視窗標題：Prompt 編輯器
+
+        # --- Layouts (版面配置) ---
+        main_layout = QVBoxLayout(self)
+        top_layout = QHBoxLayout()
+        editor_layout = QHBoxLayout()
+        bottom_layout = QHBoxLayout()
+
+        # --- ComboBox for selecting files (下拉選單選擇檔案) ---
+        self.file_selector = QComboBox()
+        self.file_selector.addItems(self.prompt_files)
+        self.file_selector.currentIndexChanged.connect(self.load_file_content)
+
+        # --- Left Editable panel (左邊編輯區塊) ---
+        self.middle_editor = QTextEdit()
+        prompt_group = QGroupBox(language_wrapper.language_word_dict.get(
+            "cot_prompt_editor_groupbox_edit_file_content"
+        ))  # 左邊編輯檔案內容
+        middle_layout = QVBoxLayout()
+        middle_layout.addWidget(self.middle_editor)
+        prompt_group.setLayout(middle_layout)
+
+        editor_layout.addWidget(prompt_group, 1)
+
+        # --- Buttons ---
+        self.create_button = QPushButton(language_wrapper.language_word_dict.get(
+            "cot_prompt_editor_button_create_file"
+        ))
+        self.create_button.clicked.connect(self.create_file)
+
+        self.save_button = QPushButton(language_wrapper.language_word_dict.get(
+            "cot_prompt_editor_button_save_file"
+        ))
+        self.save_button.clicked.connect(self.save_file)
+
+        self.reload_button = QPushButton(language_wrapper.language_word_dict.get(
+            "cot_prompt_editor_button_reload_file"
+        ))
+        self.reload_button.clicked.connect(lambda: self.load_file_content(self.file_selector.currentIndex()))
+
+        bottom_layout.addWidget(self.file_selector)
+        bottom_layout.addStretch()
+        bottom_layout.addWidget(self.reload_button)
+        bottom_layout.addWidget(self.save_button)
+        bottom_layout.addWidget(self.create_button)
+
+        # --- Combine layouts (組合版面配置) ---
+        main_layout.addLayout(top_layout)
+        main_layout.addLayout(editor_layout)
+        main_layout.addLayout(bottom_layout)
+
+        # --- FileSystemWatcher (檔案監控器) ---
+        self.watcher = QFileSystemWatcher(self.prompt_files)
+        self.watcher.fileChanged.connect(self.on_file_changed)
+
+        # 預設載入第一個檔案
+        self.load_file_content(0)
+
+    def load_file_content(self, index):
+        """載入選擇的檔案內容到左邊編輯區"""
+        filename = self.prompt_files[index]
+        self.current_file = filename
+        if os.path.exists(filename):
+            with open(filename, "r", encoding="utf-8") as f:
+                content = f.read()
+            self.middle_editor.setPlainText(content)
+        else:
+            self.middle_editor.setPlainText(language_wrapper.language_word_dict.get(
+                "cot_prompt_editor_file_not_exist"
+            ).format(filename=filename))
+
+    def create_file(self):
+        """建立目前選擇的檔案，若不存在則用模板內容建立"""
+        filename = self.current_file
+        if os.path.exists(filename):
+            QMessageBox.information(
+                self,
+                language_wrapper.language_word_dict.get("cot_prompt_editor_msgbox_info_title"),
+                language_wrapper.language_word_dict.get("cot_prompt_editor_msgbox_file_exists").format(filename=filename))
+            return
+
+        template_content = self.templates.get(filename, "")
+        with open(filename, "w", encoding="utf-8") as f:
+            f.write(template_content)
+
+        QMessageBox.information(
+            self,
+            language_wrapper.language_word_dict.get("cot_prompt_editor_msgbox_success_title"),
+            language_wrapper.language_word_dict.get("cot_prompt_editor_msgbox_file_created").format(filename=filename))
+        self.load_file_content(self.file_selector.currentIndex())
+
+    def on_file_changed(self, path):
+        """當檔案被外部修改時即時更新"""
+        if path == self.current_file:
+            self.load_file_content(self.file_selector.currentIndex())
+
+    def save_file(self):
+        """將左邊編輯區內容儲存到目前檔案"""
+        if not hasattr(self, "current_file"):
+            QMessageBox.warning(
+                self,
+                language_wrapper.language_word_dict.get("cot_prompt_editor_msgbox_error_title"),
+                language_wrapper.language_word_dict.get("cot_cot_prompt_editor_msgbox_no_file_selected"))
+            return
+
+        content = self.middle_editor.toPlainText()
+        with open(self.current_file, "w", encoding="utf-8") as f:
+            f.write(content)
+        QMessageBox.information(
+            self,
+            language_wrapper.language_word_dict.get("cot_prompt_editor_msgbox_success_title"),
+            language_wrapper.language_word_dict.get("cot_prompt_editor_msgbox_file_saved").format(
+                filename=self.current_file))

automation_ide/automation_editor_ui/extend_ai_gui/prompt_edit_gui/skills_prompt_editor_widget.py

@@ -0,0 +1,136 @@
+import os
+
+from PySide6.QtCore import QFileSystemWatcher
+from PySide6.QtWidgets import (
+    QWidget, QVBoxLayout, QHBoxLayout,
+    QComboBox, QTextEdit, QPushButton, QGroupBox, QMessageBox
+)
+from je_editor import language_wrapper
+
+from automation_ide.automation_editor_ui.extend_ai_gui.ai_gui_global_variable import SKILLS_TEMPLATE_FILES, \
+    SKILLS_TEMPLATE_RELATION
+
+
+class SkillPromptEditor(QWidget):
+    def __init__(self, skill_files=None, parent=None):
+        super().__init__(parent)
+        self.skill_files = skill_files or SKILLS_TEMPLATE_FILES
+
+        # 對應檔案名稱與模板內容
+        self.templates = SKILLS_TEMPLATE_RELATION
+
+        self.setWindowTitle(language_wrapper.language_word_dict.get(
+            "skill_prompt_editor_window_title"
+        ))  # 視窗標題：Skill Prompt 編輯器
+
+        # --- Layouts ---
+        main_layout = QVBoxLayout(self)
+        editor_layout = QHBoxLayout()
+        bottom_layout = QHBoxLayout()
+
+        # --- ComboBox for selecting files ---
+        self.file_selector = QComboBox()
+        self.file_selector.addItems(self.skill_files)
+        self.file_selector.currentIndexChanged.connect(self.load_file_content)
+
+        # --- Editable panel ---
+        self.middle_editor = QTextEdit()
+        skill_group = QGroupBox(language_wrapper.language_word_dict.get(
+            "skill_prompt_editor_groupbox_edit_file_content"
+        ))
+        middle_layout = QVBoxLayout()
+        middle_layout.addWidget(self.middle_editor)
+        skill_group.setLayout(middle_layout)
+
+        editor_layout.addWidget(skill_group, 1)
+
+        # --- Buttons ---
+        self.create_button = QPushButton(language_wrapper.language_word_dict.get(
+            "skill_prompt_editor_button_create_file"
+        ))
+        self.create_button.clicked.connect(self.create_file)
+
+        self.save_button = QPushButton(language_wrapper.language_word_dict.get(
+            "skill_prompt_editor_button_save_file"
+        ))
+        self.save_button.clicked.connect(self.save_file)
+
+        self.reload_button = QPushButton(language_wrapper.language_word_dict.get(
+            "skill_prompt_editor_button_reload_file"
+        ))
+        self.reload_button.clicked.connect(lambda: self.load_file_content(self.file_selector.currentIndex()))
+
+        bottom_layout.addWidget(self.file_selector)
+        bottom_layout.addStretch()
+        bottom_layout.addWidget(self.reload_button)
+        bottom_layout.addWidget(self.save_button)
+        bottom_layout.addWidget(self.create_button)
+
+        # --- Combine layouts ---
+        main_layout.addLayout(editor_layout)
+        main_layout.addLayout(bottom_layout)
+
+        # --- FileSystemWatcher ---
+        self.watcher = QFileSystemWatcher(self.skill_files)
+        self.watcher.fileChanged.connect(self.on_file_changed)
+
+        # 預設載入第一個檔案
+        self.load_file_content(0)
+
+    def load_file_content(self, index):
+        """載入選擇的檔案內容到編輯區"""
+        filename = self.skill_files[index]
+        self.current_file = filename
+        if os.path.exists(filename):
+            with open(filename, "r", encoding="utf-8") as f:
+                content = f.read()
+            self.middle_editor.setPlainText(content)
+        else:
+            self.middle_editor.setPlainText(language_wrapper.language_word_dict.get(
+                "skill_prompt_editor_file_not_exist"
+            ).format(filename=filename))
+
+    def create_file(self):
+        """建立目前選擇的檔案，若不存在則用模板內容建立"""
+        filename = self.current_file
+        if os.path.exists(filename):
+            QMessageBox.information(
+                self,
+                language_wrapper.language_word_dict.get("skill_prompt_editor_msgbox_info_title"),
+                language_wrapper.language_word_dict.get("skill_prompt_editor_msgbox_file_exists").format(
+                    filename=filename))
+            return
+
+        template_content = self.templates.get(filename, "")
+        with open(filename, "w", encoding="utf-8") as f:
+            f.write(template_content)
+
+        QMessageBox.information(
+            self,
+            language_wrapper.language_word_dict.get("skill_prompt_editor_msgbox_success_title"),
+            language_wrapper.language_word_dict.get("skill_prompt_editor_msgbox_file_created").format(
+                filename=filename))
+        self.load_file_content(self.file_selector.currentIndex())
+
+    def on_file_changed(self, path):
+        """當檔案被外部修改時即時更新"""
+        if path == self.current_file:
+            self.load_file_content(self.file_selector.currentIndex())
+
+    def save_file(self):
+        """將編輯區內容儲存到目前檔案"""
+        if not hasattr(self, "current_file"):
+            QMessageBox.warning(
+                self,
+                language_wrapper.language_word_dict.get("skill_prompt_editor_msgbox_error_title"),
+                language_wrapper.language_word_dict.get("skill_prompt_editor_msgbox_no_file_selected"))
+            return
+
+        content = self.middle_editor.toPlainText()
+        with open(self.current_file, "w", encoding="utf-8") as f:
+            f.write(content)
+        QMessageBox.information(
+            self,
+            language_wrapper.language_word_dict.get("skill_prompt_editor_msgbox_success_title"),
+            language_wrapper.language_word_dict.get("skill_prompt_editor_msgbox_file_saved").format(
+                filename=self.current_file))

automation_ide/automation_editor_ui/extend_ai_gui/prompt_edit_gui/skills_prompt_templates/code_explainer.py

@@ -0,0 +1,21 @@
+CODE_EXPLAINER_TEMPLATE = """
+You are a senior software engineer and educator. Explain the following code clearly and in detail.
+
+Focus on:
+- What the code does (purpose).
+- How it works (step-by-step flow, inputs/outputs).
+- Key functions, classes, or modules.
+- Assumptions, edge cases, and possible errors.
+- Performance or security concerns.
+- Suggested improvements.
+
+Format:
+- Title: one-line summary.
+- Overview: short description.
+- Detailed explanation: bullet points for flow and components.
+- Improvements: concise list with rationale.
+- Example usage if relevant.
+
+Code:
+{code_diff}
+"""

automation_ide/automation_editor_ui/extend_ai_gui/prompt_edit_gui/skills_prompt_templates/code_review.py

@@ -0,0 +1,37 @@
+CODE_REVIEW_SKILL_TEMPLATE = """
+You are an expert software reviewer. Your task is to analyze one or more code diffs from a Pull Request (PR).  
+The input may contain multiple `code_diff` sections placed in different positions.  
+
+For each diff, perform the following analysis and **always output results in three distinct sections**:
+
+---
+
+### 1. **Summary**
+- Explain the purpose and scope of the changes.  
+- Highlight affected files, functions, or modules.  
+- Provide a plain-language explanation for non-experts.  
+
+---
+
+### 2. **Linting Issues**
+- Check the code against style and formatting rules.  
+- List violations clearly, with file and line references if possible.  
+- Suggest corrections following best practices.  
+
+---
+
+### 3. **Code Smells**
+- Identify potential maintainability issues (duplication, long functions, poor naming, tight coupling).  
+- Explain why each issue is problematic.  
+- Recommend improvements or refactoring strategies.  
+
+---
+
+## Code diff
+{code_diff}
+
+### Output Format
+- Group results by diff: e.g., **Diff #1**, **Diff #2**, etc.  
+- Within each diff, strictly divide output into the three sections: **Summary**, **Linting Issues**, **Code Smells**.  
+- Present findings in bullet points or tables for readability. 
+"""