diff --git a/package-lock.json b/package-lock.json
index 799427f..ba4a88d 100644
--- a/package-lock.json
+++ b/package-lock.json
@@ -1,12 +1,12 @@
 {
   "name": "g1nation",
-  "version": "2.72.0",
+  "version": "2.73.0",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "g1nation",
-      "version": "2.72.0",
+      "version": "2.73.0",
       "license": "MIT",
       "dependencies": {
         "marked": "^18.0.2"
diff --git a/package.json b/package.json
index 9641c5f..778e8a7 100644
--- a/package.json
+++ b/package.json
@@ -2,7 +2,7 @@
   "name": "astra",
   "displayName": "Astra",
   "description": "The personal intelligence layer for Antigravity and VS Code. A private cognitive partner for deep project context, memory, and proactive strategic decision-making.",
-  "version": "2.72.0",
+  "version": "2.73.0",
   "publisher": "g1nation",
   "license": "MIT",
   "icon": "assets/icon.png",
diff --git a/src/retrieval/scoring.ts b/src/retrieval/scoring.ts
index 47b77fa..90378d8 100644
--- a/src/retrieval/scoring.ts
+++ b/src/retrieval/scoring.ts
@@ -168,6 +168,8 @@ function inverseDocumentFrequency(
     return Math.log((smoothN + 1) / (smoothContaining + 1)) + 1;
 }
 
+export type ConflictSeverity = 'NONE' | 'LOW' | 'MEDIUM' | 'HIGH';
+
 export interface ScoredDocument {
     index: number;
     score: number;
@@ -175,6 +177,7 @@ export interface ScoredDocument {
     recencyBoost: number;
     matchedTerms: string[];
     conflictDetected: boolean;
+    conflictSeverity: ConflictSeverity;
     informationDensity: number;
 }
 
@@ -216,8 +219,18 @@ export function scoreTfIdf(
         let score = 0;
         const matchedTerms: string[] = [];
 
-        // Conflict Detection: 문서 내 상충 지표 확인
-        const conflictDetected = docTokens.some(t => SCORING_CONFIG.CONFLICT_INDICATORS.has(t));
+        // Conflict Detection & Severity Analysis (Substring based for better recall with particles)
+        const rawText = `${doc.title} ${doc.content}`.toLowerCase();
+        const conflictMatches = [...SCORING_CONFIG.CONFLICT_INDICATORS].filter(indicator => 
+            rawText.includes(indicator.toLowerCase())
+        );
+        
+        const conflictDetected = conflictMatches.length > 0;
+        let conflictSeverity: ConflictSeverity = 'NONE';
+
+        if (conflictMatches.length >= 4) conflictSeverity = 'HIGH';
+        else if (conflictMatches.length >= 2) conflictSeverity = 'MEDIUM';
+        else if (conflictMatches.length === 1) conflictSeverity = 'LOW';
 
         for (const term of expandedQuery) {
             const tf = termFrequency(term, docTokens);
@@ -255,6 +268,7 @@ export function scoreTfIdf(
             recencyBoost,
             matchedTerms: [...new Set(matchedTerms)],
             conflictDetected,
+            conflictSeverity,
             informationDensity
         };
     });
diff --git a/tests/scoring.test.ts b/tests/scoring.test.ts
index a7c3792..88f2535 100644
--- a/tests/scoring.test.ts
+++ b/tests/scoring.test.ts
@@ -28,17 +28,19 @@ describe('Scoring Engine Unit Tests (v2.72.0)', () => {
         expect(expanded).toContain('optimization');
     });
 
-    test('Conflict Detection: should flag documents with controversial terms', () => {
+    test('Conflict Detection & Severity: should flag documents with tiered severity', () => {
         const query = ['설계'];
         const docs = [
             { title: '정상 설계 문서', content: '이 시스템은 효율적으로 설계되었습니다.' },
-            { title: '상충 발생 문서', content: '이 설계는 기존 아키텍처와 충돌 논란이 있습니다.' }
+            { title: '상충 발생 문서 (LOW)', content: '이 설계는 기존 아키텍처와 충돌 위험이 있습니다.' },
+            { title: '강한 상충 문서 (HIGH)', content: '이 설계는 오류가 많고 논란이 크며 반대 의견과 반박이 거셉니다.' }
         ];
         
         const results = scoreTfIdf(tokenize(query.join(' ')), docs);
         
-        expect(results[0].conflictDetected).toBe(false);
-        expect(results[1].conflictDetected).toBe(true);
+        expect(results[0].conflictSeverity).toBe('NONE');
+        expect(results[1].conflictSeverity).toBe('LOW');
+        expect(results[2].conflictSeverity).toBe('HIGH');
     });
 
     test('IDF Smoothing: should provide stable scores for small datasets', () => {