chore: version bump to 2.80.9 and fix bridge/lmstudio issues

.astra/tests/stress/.astra/cache/259a37934ead3910a8722b82054d46d2ca2057b05c488be1dcf439166ac5a9a1.json

@@ -1,5 +1,5 @@
 {
   "result": "Final report with inconsistencies. This should be long enough to pass validation.",
-  "createdAt": 1778045762198,
+  "createdAt": 1778129530764,
   "modelVersion": "unknown"
 }

.astra/tests/stress/.astra/cache/65775be352df43297b63c7af59c9f4f39d2bc368f77456c37b5eef9a94a66b5c.json

@@ -1,5 +1,5 @@
 {
   "result": "[CONFLICT WARNING] 성능이 200% 증가했습니다. vs 그러나 동시에 50% 감소했습니다. 최적화와 성능 저하가 동시에 발견됨.",
-  "createdAt": 1778045762196,
+  "createdAt": 1778129530762,
   "modelVersion": "unknown"
 }

.astra/tests/stress/.astra/cache/6894d26c5b0a55d25d756a473225c7a44d7661af673b24e3f49551a7a2e50280.json

@@ -1,5 +1,5 @@
 {
   "result": "Detailed Execution Plan: 1. Research 2. Analyze 3. Write report with high quality.",
-  "createdAt": 1778045762195,
+  "createdAt": 1778129530759,
   "modelVersion": "unknown"
 }

.astra/tests/stress/.astra/cache/88cb61499f88ed38165b64bd3e8adc543795e4b427b64540a49c9ab27c7fe213.json

@@ -1,5 +1,5 @@
 {
-  "result": "---\nid: stress_conflict_1778045762177\ndate: 2026-05-06T05:36:02.199Z\ntype: knowledge_artifact\nstandard: P-Reinforce v3.0\ntags: [automated, connect_ai, brain_sync]\n---\n\n## 📌 Brief Summary\nFinal report with inconsistencies. This should be long enough to pass validation.\n\nFinal report with inconsistencies. This should be long enough to pass validation.\n\n---\n## 💡 Astra의 선제적 제안 (Proactive Next Actions)\nFinal report with inconsistencies. This should be long enough to pass validation.\n---\n## 🛡️ Reliability & Audit Summary\n> [!NOTE]\n> 이 문서는 ConnectAI의 **Intelligent Resilience** 엔진에 의해 검증 및 정제되었습니다.\n\n| Metric | Value | Status |\n| :--- | :--- | :--- |\n| **Conflict Risk** | `60/100` | ⚠️ Medium |\n| **Fallbacks Used** | `0` | ✅ None |\n| **Auto Retries** | `0` | ✅ Stable |\n| **Deduplication** | `0` | Standard |\n| **Processing Time** | `0.0s` | ✅ Fast |\n\n### 🔍 Decision Audit Trail\n- **[PLANNER]** 전략 수립 중... (16ms)\n- **[RESEARCHER]** 핵심 정보 수집 및 분석 중... (2ms)\n- **[WRITER]** 최종 리포트 작성 및 편집 중... (3ms)\n",
+  "result": "---\nid: stress_conflict_1778129530743\ndate: 2026-05-07T04:52:10.766Z\ntype: knowledge_artifact\nstandard: P-Reinforce v3.0\ntags: [automated, connect_ai, brain_sync]\n---\n\n## 📌 Brief Summary\nFinal report with inconsistencies. This should be long enough to pass validation.\n\nFinal report with inconsistencies. This should be long enough to pass validation.\n\n---\n## 💡 Astra의 선제적 제안 (Proactive Next Actions)\nFinal report with inconsistencies. This should be long enough to pass validation.\n---\n## 🛡️ Reliability & Audit Summary\n> [!NOTE]\n> 이 문서는 ConnectAI의 **Intelligent Resilience** 엔진에 의해 검증 및 정제되었습니다.\n\n| Metric | Value | Status |\n| :--- | :--- | :--- |\n| **Conflict Risk** | `60/100` | ⚠️ Medium |\n| **Fallbacks Used** | `0` | ✅ None |\n| **Auto Retries** | `0` | ✅ Stable |\n| **Deduplication** | `0` | Standard |\n| **Processing Time** | `0.0s` | ✅ Fast |\n\n### 🔍 Decision Audit Trail\n- **[PLANNER]** 전략 수립 중... (15ms)\n- **[RESEARCHER]** 핵심 정보 수집 및 분석 중... (3ms)\n- **[WRITER]** 최종 리포트 작성 및 편집 중... (2ms)\n",
-  "createdAt": 1778045762200,
+  "createdAt": 1778129530766,
   "modelVersion": "unknown"
 }

.astra/tests/stress/.astra/missions/stress_conflict_1778129530743.json

@@ -1,8 +1,8 @@
 {
-  "missionId": "stress_conflict_1778045762177",
+  "missionId": "stress_conflict_1778129530743",
   "status": "completed",
-  "startTime": "2026-05-06T05:36:02.177Z",
+  "startTime": "2026-05-07T04:52:10.743Z",
-  "totalElapsedMs": 23,
+  "totalElapsedMs": 24,
   "results": {
     "planner": "Detailed Execution Plan: 1. Research 2. Analyze 3. Write report with high quality.",
     "researcher": "[CONFLICT WARNING] 성능이 200% 증가했습니다. vs 그러나 동시에 50% 감소했습니다. 최적화와 성능 저하가 동시에 발견됨.",
@@ -16,30 +16,30 @@
     {
       "from": "idle",
       "to": "planner",
-      "durationMs": 16,
+      "durationMs": 15,
       "message": "전략 수립 중...",
-      "ts": "2026-05-06T05:36:02.193Z"
+      "ts": "2026-05-07T04:52:10.758Z"
     },
     {
       "from": "planner",
       "to": "researcher",
-      "durationMs": 2,
+      "durationMs": 3,
       "message": "핵심 정보 수집 및 분석 중...",
-      "ts": "2026-05-06T05:36:02.195Z"
+      "ts": "2026-05-07T04:52:10.761Z"
     },
     {
       "from": "researcher",
       "to": "writer",
-      "durationMs": 3,
+      "durationMs": 2,
       "message": "최종 리포트 작성 및 편집 중...",
-      "ts": "2026-05-06T05:36:02.198Z"
+      "ts": "2026-05-07T04:52:10.763Z"
     },
     {
       "from": "writer",
       "to": "completed",
-      "durationMs": 2,
+      "durationMs": 4,
       "message": "미션 완료",
-      "ts": "2026-05-06T05:36:02.200Z"
+      "ts": "2026-05-07T04:52:10.767Z"
     }
   ],
   "resilienceMetrics": {

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": "2.80.5",
+  "version": "2.80.9",
   "publisher": "g1nation",
   "license": "MIT",
   "icon": "assets/icon.png",

src/agent.ts

@@ -451,35 +451,74 @@ export class AgentExecutor {
             if (this.isStaleRun(runId)) return;
 
             let aiResponseText = '';
-            const reader = response.body?.getReader();
+            const body = response.body as any;
-            if (!reader) throw new Error("Response body is not readable.");
+            if (!body) throw new Error("Response body is null.");
 
-            if (loopDepth === 0) this.webview.postMessage({ type: 'streamStart' });
+            if (loopDepth === 0) this.webview?.postMessage({ type: 'streamStart' });
 
             let buffer = '';
             const decoder = new TextDecoder();
-            try {
-                while (true) {
-                    const { done, value } = await reader.read();
-                    if (done) break;
-                    if (this.isStaleRun(runId)) return;
 
-                    buffer += decoder.decode(value, { stream: true });
+            const processChunk = (value: any) => {
-                    const lines = buffer.split('\n');
+                if (this.isStaleRun(runId)) return false;
-                    buffer = lines.pop() || '';
+
-                    for (const line of lines) {
+                buffer += decoder.decode(value, { stream: true });
-                        const trimmed = line.trim();
+                const lines = buffer.split('\n');
-                        if (!trimmed || trimmed === 'data: [DONE]') continue;
+                buffer = lines.pop() || '';
-                        try {
+
-                            const raw = trimmed.startsWith('data: ') ? trimmed.slice(6) : trimmed;
+                for (const line of lines) {
-                            const json = JSON.parse(raw);
+                    const trimmed = line.trim();
-                            const token = engine === 'lmstudio' ? json.choices?.[0]?.delta?.content || '' : json.message?.content || json.response || '';
+                    if (!trimmed || trimmed === 'data: [DONE]') continue;
-                            if (token) {
+                    
-                                aiResponseText += token;
+                    try {
-                            }
+                        let raw = trimmed;
-                        } catch (e: any) {
+                        if (trimmed.startsWith('data:')) {
-                            logError('Failed to parse streaming chunk.', { engine, apiUrl, chunk: summarizeText(trimmed, 300), error: e?.message || String(e) });
+                            raw = trimmed.replace(/^data:\s*/, '');
                         }
+                        
+                        if (!raw || raw === '[DONE]') continue;
+                        
+                        const json = JSON.parse(raw);
+                        if (json.error) {
+                            const errMsg = typeof json.error === 'string' ? json.error : (json.error.message || JSON.stringify(json.error));
+                            throw new Error(`AI Engine Error: ${errMsg}`);
+                        }
+                        
+                        let token = '';
+                        if (json.choices?.[0]) {
+                            const choice = json.choices[0];
+                            token = choice.delta?.content || choice.message?.content || choice.text || '';
+                        } else if (json.message?.content) {
+                            token = json.message.content;
+                        } else if (json.response) {
+                            token = json.response;
+                        }
+                        
+                        if (token) {
+                            aiResponseText += token;
+                            if (loopDepth === 0) {
+                                this.webview?.postMessage({ type: 'streamUpdate', value: token });
+                            }
+                        }
+                    } catch (e: any) {
+                        // Silent fail for non-JSON lines unless it's an AI Engine Error
+                        if (e.message.startsWith('AI Engine Error:')) throw e;
+                    }
+                }
+                return true;
+            };
+
+            try {
+                if (typeof body[Symbol.asyncIterator] === 'function') {
+                    for await (const chunk of body) {
+                        if (!processChunk(chunk)) break;
+                    }
+                } else {
+                    const reader = body.getReader();
+                    while (true) {
+                        const { done, value } = await reader.read();
+                        if (done) break;
+                        if (!processChunk(value)) break;
                     }
                 }
             } catch (err: any) {
@@ -495,11 +534,30 @@ export class AgentExecutor {
             if (buffer.trim() && buffer.trim() !== 'data: [DONE]') {
                 try {
                     const trimmed = buffer.trim();
-                    const raw = trimmed.startsWith('data: ') ? trimmed.slice(6) : trimmed;
+                    let raw = trimmed;
-                    const json = JSON.parse(raw);
+                    if (trimmed.startsWith('data:')) {
-                    const token = engine === 'lmstudio' ? json.choices?.[0]?.delta?.content || '' : json.message?.content || json.response || '';
+                        raw = trimmed.replace(/^data:\s*/, '');
-                    if (token) {
+                    }
-                        aiResponseText += token;
+                    
+                    if (raw && raw !== '[DONE]') {
+                        const json = JSON.parse(raw);
+                        if (json.error) {
+                            const errMsg = typeof json.error === 'string' ? json.error : (json.error.message || JSON.stringify(json.error));
+                            throw new Error(`AI Engine Error: ${errMsg}`);
+                        }
+                        let token = '';
+                        if (json.choices?.[0]) {
+                            const choice = json.choices[0];
+                            token = choice.delta?.content || choice.message?.content || choice.text || '';
+                        } else if (json.message?.content) {
+                            token = json.message.content;
+                        } else if (json.response) {
+                            token = json.response;
+                        }
+                        
+                        if (token) {
+                            aiResponseText += token;
+                        }
                     }
                 } catch (e: any) {
                     logError('Failed to parse final streaming buffer.', { engine, apiUrl, buffer: summarizeText(buffer, 300), error: e?.message || String(e) });
@@ -556,24 +614,36 @@ export class AgentExecutor {
             */
             const finalAssistantContent = assistantContent;
 
+            const assistantMessage: ChatMessage = { role: 'assistant', content: finalAssistantContent, internal: false, rationale };
+            this.chatHistory.push(assistantMessage);
+            this.emitHistoryChanged();
+
             this.statusBarManager.updateStatus(AgentStatus.Executing);
             const report = await this.executeActions(aiResponseText, rootPath, activeBrain);
             if (!assistantContent.trim() && report.length === 0) {
-                logError('Model returned an empty response without actions.', { model: actualModel, engine, apiUrl, loopDepth });
+                const totalChars2 = messagesForRequest.reduce((acc, m) => acc + String(m.content || '').length, 0);
+                const estimatedTokens2 = Math.ceil(totalChars2 / 4);
+                const isContextOverflow = estimatedTokens2 > 5000;
+                logError('Model returned an empty response without actions.', { model: actualModel, engine, apiUrl, loopDepth, estimatedTokens: estimatedTokens2 });
                 this.webview.postMessage({
                     type: 'error',
                     value: [
                         'AI engine returned an empty response.',
-                        `Engine: ${engine}`,
+                        `Engine: ${engine}  |  Model: ${actualModel}`,
-                        `Model: ${actualModel}`,
+                        isContextOverflow
-                        'The request reached the local LLM server, but no usable content was returned. Try another model, restart the local server, or reduce the prompt/context size.'
+                            ? `Context overflow: ~${estimatedTokens2.toLocaleString()} tokens estimated. This model likely has a smaller context window.`
+                            : 'The request reached the LLM server, but no content was returned.',
+                        '',
+                        '**해결 방법:**',
+                        isContextOverflow
+                            ? '1. Brain 비활성화 후 재시도  2. 더 큰 모델(7B+) 사용  3. 대화 기록 초기화 후 재시도'
+                            : '1. LM Studio에서 해당 모델이 로드되어 있는지 확인  2. 모델 재시작 후 재시도  3. 다른 모델로 전환'
                     ].join('\n')
                 });
                 return;
             }
 
             if (report.length > 0) {
-                this.emitHistoryChanged();
                 logInfo('Agent actions executed.', { loopDepth: loopDepth + 1, report });
 
                 // Continue loop if needed
@@ -590,7 +660,7 @@ export class AgentExecutor {
                     logInfo('Autonomous loop continuing after actions.', { loopDepth: loopDepth + 1, actions: report });
 
                     // Explicitly tell the AI to look at the results and continue
-                    const continuationPrompt = "The requested local action has been executed. Use the action result messages already in the conversation to answer the user's original request directly, in the user's language. Do not say you are waiting for the next instruction.";
+                    const continuationPrompt = `The requested local action has been executed.\nAction report:\n${report.join('\n')}\nUse the action result messages already in the conversation to answer the user's original request directly, in the user's language. Do not say you are waiting for the next instruction.`;
 
                     this.webview.postMessage({ type: 'autoContinue', value: `자료를 확인하고 답변을 정리하는 중입니다... (${loopDepth + 1}/${config.maxAutoSteps})` });
                     await new Promise(r => setTimeout(r, 800));
@@ -600,9 +670,6 @@ export class AgentExecutor {
                 return;
             }
 
-            const assistantMessage: ChatMessage = { role: 'assistant', content: finalAssistantContent, internal: false, rationale };
-            this.chatHistory.push(assistantMessage);
-            this.emitHistoryChanged();
             this.statusBarManager.updateStatus(AgentStatus.Success);
             this.webview.postMessage({ type: 'streamChunk', value: finalAssistantContent });
 
@@ -1953,14 +2020,29 @@ export class AgentExecutor {
 
             for (const candidateModel of modelCandidates) {
                 for (const variant of messageVariants) {
+                    // LM Studio: context_length를 명시적으로 제한하여 컨텍스트 초과 방지
+                    // 총 메시지 토큰 추정: 문자 수 / 4 (rough estimate)
+                    const totalChars = variant.messages.reduce((acc, m) => acc + String(m.content || '').length, 0);
+                    const estimatedTokens = Math.ceil(totalChars / 4);
+                    // LM Studio 소형 모델(4B~8B)은 4096~8192 context 제한
+                    // 컨텍스트 초과 시 max_tokens을 줄여서 모델이 응답할 공간 확보
+                    const lmStudioMaxTokens = Math.max(512, Math.min(4096, 8192 - estimatedTokens));
                     const streamBody = {
                         model: candidateModel,
                         messages: variant.messages,
                         stream: true,
                         ...(engine === 'lmstudio'
-                            ? { max_tokens: 4096, temperature }
+                            ? {
+                                max_tokens: lmStudioMaxTokens,
+                                temperature,
+                                // LM Studio: context_length로 컨텍스트 창 명시 설정
+                                context_length: 8192
+                            }
                             : { options: { num_ctx: 32768, num_predict: 4096, temperature } }),
                     };
+                    if (engine === 'lmstudio' && estimatedTokens > 6000) {
+                        logError('LM Studio context may be too large for small models.', { estimatedTokens, lmStudioMaxTokens, model: candidateModel });
+                    }
 
                     try {
                         logInfo('AI streaming request started.', {
@@ -1981,8 +2063,7 @@ export class AgentExecutor {
                                 'Connection': 'keep-alive'
                             },
                             body: JSON.stringify(streamBody),
-                            signal: this.abortController?.signal,
+                            signal: this.abortController?.signal
-                            keepalive: true
                         });
 
                         if (!response.ok) {

src/bridge.ts

@@ -39,7 +39,7 @@ export class BridgeServer {
     }
 
     public start(port: number = 4825) {
-        this.server = http.createServer((req, res) => {
+        const server = http.createServer((req, res) => {
             res.setHeader('Access-Control-Allow-Origin', '*');
             res.setHeader('Access-Control-Allow-Methods', 'GET, POST, OPTIONS');
             res.setHeader('Access-Control-Allow-Headers', 'Content-Type');
@@ -70,15 +70,24 @@ export class BridgeServer {
             }
         });
 
-        this.server.on('error', (err: any) => {
+        // once() 사용: 중복 에러 이벤트 방지
+        server.once('error', (err: any) => {
             if (err.code === 'EADDRINUSE') {
-                logError(`🚫 Bridge Port ${port} in use. Connection with EZER/A.U might fail.`);
+                logInfo(`Bridge Port ${port} already in use. Trying port ${port + 1}...`);
+                // 기존 서버 참조 정리 후 다음 포트 시도
+                server.close();
+                if (this.server === server) {
+                    this.server = null;
+                }
+                this.start(port + 1);
             } else {
-                logError(`Bridge server error:`, err);
+                logError(`Bridge server error on port ${port}:`, err);
             }
         });
 
-        this.server.listen(port, '127.0.0.1', () => {
+        // 성공 시 서버 참조 저장
+        server.listen(port, '127.0.0.1', () => {
+            this.server = server;
             logInfo(`Bridge server active on 127.0.0.1:${port}.`);
         });
     }