Add review risk assessment and optional repository review rules

pr_agent/algo/utils.py

@@ -236,6 +236,38 @@ def convert_to_markdown_v2(output_data: dict,
                     markdown_text += f"### {emoji} Security concerns\n\n"
                     value = emphasize_header(value.strip(), only_markdown=True)
                     markdown_text += f"{value}\n\n"
+        elif 'risk level' in key_nice.lower():
+            risk_value = str(value).strip().lower().replace('_', ' ')
+            risk_display = risk_value.capitalize() if risk_value else 'Unknown'
+            if gfm_supported:
+                markdown_text += f"<tr><td><strong>Risk level</strong>: {risk_display}</td></tr>\n"
+            else:
+                markdown_text += f"### Risk level: {risk_display}\n\n"
+        elif 'merge recommendation' in key_nice.lower():
+            recommendation = str(value).strip().replace('_', ' ')
+            recommendation_display = recommendation.capitalize() if recommendation else 'Unknown'
+            if gfm_supported:
+                markdown_text += f"<tr><td><strong>Merge recommendation</strong>: {recommendation_display}</td></tr>\n"
+            else:
+                markdown_text += f"### Merge recommendation: {recommendation_display}\n\n"
+        elif 'review priority files' in key_nice.lower():
+            if gfm_supported:
+                markdown_text += "<tr><td>"
+                if not value:
+                    markdown_text += "<strong>Priority files</strong>: None"
+                else:
+                    markdown_text += "<strong>Priority files</strong><br><br>"
+                    for priority_file in value:
+                        markdown_text += f"- {priority_file}<br>"
+                markdown_text += "</td></tr>\n"
+            else:
+                if not value:
+                    markdown_text += "### Priority files: None\n\n"
+                else:
+                    markdown_text += "### Priority files\n\n"
+                    for priority_file in value:
+                        markdown_text += f"- {priority_file}\n"
+                    markdown_text += "\n"
         elif 'todo sections' in key_nice.lower():
             if gfm_supported:
                 markdown_text += "<tr><td>"

pr_agent/settings/configuration.toml

@@ -81,6 +81,12 @@ require_security_review=true
 require_estimate_contribution_time_cost=false
 require_todo_scan=false
 require_ticket_analysis_review=true
+require_risk_assessment=true
+require_merge_recommendation=true
+require_priority_files=true
+enable_review_rules=true
+review_rules_paths=[".pr_agent/review_rules.md",".github/review_rules.md","docs/review_rules.md"]
+max_review_rules_tokens=1200
 # general options
 publish_output_no_suggestions=true # Set to "false" if you only need the reviewer's remarks (not labels, not "security audit", etc.) and want to avoid noisy "No major issues detected" comments.
 persistent_comment=true

pr_agent/settings/pr_reviewer_prompts.toml

@@ -68,7 +68,16 @@ Extra instructions from the user:
 {{ extra_instructions }}
 ======
 {% endif %}
+{%- if review_rules %}
 
+Repository/team review rules:
+======
+{{ review_rules }}
+======
+- Apply these repository rules in addition to the general review criteria.
+- When a repository rule conflicts with a generic style preference, prioritize the repository rule.
+- Use the repository rules to decide what deserves extra attention, but do not invent violations that are not supported by the diff.
+{% endif %}
 
 The output must be a YAML object equivalent to type $PRReview, according to the following Pydantic definitions:
 =====
@@ -117,6 +126,15 @@ class Review(BaseModel):
 {%- if require_estimate_effort_to_review %}
     estimated_effort_to_review_[1-5]: int = Field(description="Estimate, on a scale of 1-5 (inclusive), the time and effort required to review this PR by an experienced and knowledgeable developer. 1 means short and easy review, 5 means long and hard review. Take into account the size, complexity, quality, and the needed changes of the PR code diff.")
 {%- endif %}
+{%- if require_risk_assessment %}
+    risk_level: str = Field(description="Overall risk level of this PR. Answer with exactly one of: low, medium, high. Use high only when the PR introduces a clear bug, security concern, or major logic risk. Use medium when the PR is not clearly broken but contains non-trivial areas that require careful human verification. Use low when the PR is small, low-impact, and no important issues are identified.")
+{%- endif %}
+{%- if require_merge_recommendation %}
+    merge_recommendation: str = Field(description="Overall merge recommendation for this PR. Answer with exactly one of: safe_to_merge, merge_with_caution, changes_required. Use changes_required when there are clear issues that should be fixed before merge. Use merge_with_caution when the PR seems acceptable but still deserves focused reviewer attention. Use safe_to_merge when no important blockers or risks are identified.")
+{%- endif %}
+{%- if require_priority_files %}
+    review_priority_files: List[str] = Field(description="A short list of the most important files a human reviewer should inspect first. Return an empty list if the PR is too small or no file deserves special attention.")
+{%- endif %}
 {%- if require_estimate_contribution_time_cost %}
     contribution_time_cost_estimate: ContributionTimeCostEstimate = Field(description="An estimate of the time required to implement the changes, based on the quantity, quality, and complexity of the contribution, as well as the context from the PR description and commit messages.")
 {%- endif %}
@@ -165,6 +183,19 @@ review:
   estimated_effort_to_review_[1-5]: |
     3
 {%- endif %}
+{%- if require_risk_assessment %}
+  risk_level: |
+    low
+{%- endif %}
+{%- if require_merge_recommendation %}
+  merge_recommendation: |
+    safe_to_merge
+{%- endif %}
+{%- if require_priority_files %}
+  review_priority_files:
+    - |
+      src/example.py
+{%- endif %}
 {%- if require_score %}
   score: 89
 {%- endif %}
@@ -303,6 +334,19 @@ review:
   estimated_effort_to_review_[1-5]: |
     3
 {%- endif %}
+{%- if require_risk_assessment %}
+  risk_level: |
+    low
+{%- endif %}
+{%- if require_merge_recommendation %}
+  merge_recommendation: |
+    safe_to_merge
+{%- endif %}
+{%- if require_priority_files %}
+  review_priority_files:
+    - |
+      src/example.py
+{%- endif %}
 {%- if require_score %}
   score: 89
 {%- endif %}

pr_agent/tools/pr_reviewer.py

@@ -1,6 +1,9 @@
 import copy
 import datetime
 import traceback
+
+
+
 from collections import OrderedDict
 from functools import partial
 from typing import List, Tuple
@@ -13,7 +16,7 @@
                                          get_pr_diff,
                                          retry_with_fallback_models)
 from pr_agent.algo.token_handler import TokenHandler
-from pr_agent.algo.utils import (ModelType, PRReviewHeader,
+from pr_agent.algo.utils import (ModelType, PRReviewHeader, clip_tokens,
                                  convert_to_markdown_v2, github_action_output,
                                  load_yaml, show_relevant_configurations)
 from pr_agent.config_loader import get_settings
@@ -66,6 +69,8 @@ def __init__(self, pr_url: str, is_answer: bool = False, is_auto: bool = False,
         answer_str, question_str = self._get_user_answers()
         self.pr_description, self.pr_description_files = (
             self.git_provider.get_pr_description(split_changes_walkthrough=True))
+        self.review_rules = self._get_review_rules()
+
         if (self.pr_description_files and get_settings().get("config.is_auto_command", False) and
                 get_settings().get("config.enable_ai_metadata", False)):
             add_ai_metadata_to_diff_files(self.git_provider, self.pr_description_files)
@@ -85,6 +90,9 @@ def __init__(self, pr_url: str, is_answer: bool = False, is_auto: bool = False,
             "require_score": get_settings().pr_reviewer.require_score_review,
             "require_tests": get_settings().pr_reviewer.require_tests_review,
             "require_estimate_effort_to_review": get_settings().pr_reviewer.require_estimate_effort_to_review,
+            "require_risk_assessment": get_settings().pr_reviewer.get("require_risk_assessment", False),
+            "require_merge_recommendation": get_settings().pr_reviewer.get("require_merge_recommendation", False),
+            "require_priority_files": get_settings().pr_reviewer.get("require_priority_files", False),
             "require_estimate_contribution_time_cost": get_settings().pr_reviewer.require_estimate_contribution_time_cost,
             'require_can_be_split_review': get_settings().pr_reviewer.require_can_be_split_review,
             'require_security_review': get_settings().pr_reviewer.require_security_review,
@@ -99,6 +107,8 @@ def __init__(self, pr_url: str, is_answer: bool = False, is_auto: bool = False,
             "related_tickets": get_settings().get('related_tickets', []),
             'duplicate_prompt_examples': get_settings().config.get('duplicate_prompt_examples', False),
             "date": datetime.datetime.now().strftime('%Y-%m-%d'),
+            "review_rules": self.review_rules,
+
         }
 
         self.token_handler = TokenHandler(
@@ -117,6 +127,41 @@ def parse_incremental(self, args: List[str]):
         incremental = IncrementalPR(is_incremental)
         return incremental
 
+    def _get_review_rules(self) -> str:
+        if not get_settings().pr_reviewer.get("enable_review_rules", False):
+            return ""
+
+        rule_paths = get_settings().pr_reviewer.get("review_rules_paths", []) or []
+        if isinstance(rule_paths, str):
+            rule_paths = [rule_paths]
+
+        ref = getattr(getattr(self.git_provider, 'pr', None), 'head', None)
+        ref = getattr(ref, 'sha', None) or self.git_provider.get_pr_branch()
+        loaded_rules = []
+        loaded_rule_paths = []
+
+        for rule_path in rule_paths:
+            try:
+                rule_content = self.git_provider.get_pr_file_content(rule_path, ref)
+            except Exception:
+                continue
+
+            if rule_content and rule_content.strip():
+                loaded_rule_paths.append(rule_path)
+                loaded_rules.append(f"File: `{rule_path}`\n{rule_content.strip()}")
+
+        if not loaded_rules:
+            get_logger().info("No review rules file found for this PR")
+            return ""
+
+        review_rules = "\n\n---\n\n".join(loaded_rules)
+        max_tokens = get_settings().pr_reviewer.get("max_review_rules_tokens", 0)
+        if max_tokens and int(max_tokens) > 0:
+            review_rules = clip_tokens(review_rules, int(max_tokens))
+
+        get_logger().info("Loaded review rules for this PR", artifacts={"rule_files": loaded_rule_paths})
+        return review_rules
+
     async def run(self) -> None:
         try:
             if not self.git_provider.get_files():
@@ -173,7 +218,7 @@ async def run(self) -> None:
             if get_settings().pr_reviewer.persistent_comment and not self.incremental.is_incremental:
                 final_update_message = get_settings().pr_reviewer.final_update_message
                 self.git_provider.publish_persistent_comment(pr_review,
-                                                            initial_header=f"{PRReviewHeader.REGULAR.value} 🔍",
+                                                            initial_header=f"{PRReviewHeader.REGULAR.value}",
                                                             update_header=True,
                                                             final_update_message=final_update_message, )
             else:
@@ -234,14 +279,17 @@ def _prepare_pr_review(self) -> str:
         first_key = 'review'
         last_key = 'security_concerns'
         data = load_yaml(self.prediction.strip(),
-                         keys_fix_yaml=["ticket_compliance_check", "estimated_effort_to_review_[1-5]:", "security_concerns:", "key_issues_to_review:",
+                         keys_fix_yaml=["ticket_compliance_check", "estimated_effort_to_review_[1-5]:", "risk_level:", "merge_recommendation:", "review_priority_files:", "security_concerns:", "key_issues_to_review:",
                                         "relevant_file:", "relevant_line:", "suggestion:"],
                          first_key=first_key, last_key=last_key)
         github_action_output(data, 'review')
 
         if 'review' not in data:
             get_logger().exception("Failed to parse review data", artifact={"data": data})
             return ""
+        get_logger().info(f"Risk level: {data.get('review', {}).get('risk_level')}")
+        get_logger().info(f"Merge recommendation: {data.get('review', {}).get('merge_recommendation')}")
+        get_logger().info(f"Priority files: {data.get('review', {}).get('review_priority_files')}")
 
         # move data['review'] 'key_issues_to_review' key to the end of the dictionary
         if 'key_issues_to_review' in data['review']:
@@ -273,6 +321,7 @@ def _prepare_pr_review(self) -> str:
         # Add custom labels from the review prediction (effort, security)
         self.set_review_labels(data)
 
+
         if markdown_text == None or len(markdown_text) == 0:
             markdown_text = ""