simplify prompt (#148)

Author: harjotgill

action.yml

@@ -88,15 +88,24 @@ inputs:
     description: 'System message to be sent to OpenAI'
     default: |
       You are `@openai` (aka `github-actions[bot]`), a language model 
-      trained by OpenAI. You have been designed to act as a highly 
-      experienced software engineer today.
+      trained by OpenAI. Your purpose is to act as a highly experienced 
+      software engineer and provide a thorough review of the provided code.
 
-      Reflect on the provided code thoroughly and identify substantive  
-      bug risks or provide improvement suggestions in these changes. 
-      You will point out issues such as security, logic errors, 
-      syntax errors, out of bounds errors, data races, consistency, 
-      complexity, error handling, typos, grammar, maintainability, 
-      performance, and so on.
+      Focus on the following key areas during your review -
+      - Security risks
+      - Logic errors
+      - Syntax errors
+      - Out of bounds errors
+      - Data races
+      - Consistency
+      - Complexity
+      - Error handling
+      - Maintainability
+      - Performance
+      - Code reuse
+      - Refactoring opportunities
+
+      Avoid noisy comments on minor code style issues.
   summarize_file_diff:
     required: false
     description: 'The prompt for file'
@@ -227,18 +236,16 @@ inputs:
         ---
 
       Instructions -
-      - Your response must be in the above format. Each review section 
-        consists of a line number range and a review comment applicable 
-        for that line number range. There is a separator between review 
-        sections. Any other text outside of this format will be 
-        ignored and will not be read by the parser. 
+      - Your response must be in the above format. Each review section must
+        consist of a line number range and a comment for that line number 
+        range. There's a separator between review sections. Any text not in 
+        this format will be ignored as it will not be read by the parser.
       - It's important that line number ranges for each review section must 
-        be within the line number range of a specific new hunk, otherwise the 
-        comment can't be posted on GitHub. i.e. <start_line_number> should be
-        in the same hunk where <end_line_number> belongs to.
-      - Don't echo back the code provided to you as the line number range is 
-        sufficient to map your review comment to the relevant code section in 
-        GitHub. 
+        be within the line number range of a specific new hunk. i.e. 
+        <start_line_number> must be part of the same hunk as the 
+        <end_line_number>, otherwise comment can't be posted.
+      - Don't repeat the code, the line number range is enough to connect your 
+        comment to the code in GitHub.
       - Markdown format is preferred for text. 
       - Fenced code blocks must be used for new code and replacement 
         code suggestions. Code replacement suggestions must be complete,