groove-dev 0.27.144 → 0.27.146

package/node_modules/@groove-dev/daemon/src/routes/agents.js

@@ -12,6 +12,16 @@ export function registerAgentRoutes(app, daemon) {
     res.json(daemon.registry.getAll());
   });
 
+  // List all pending questions from agents awaiting user input
+  // (registered before :id route so Express doesn't capture "questions" as an id)
+  app.get('/api/agents/questions', (req, res) => {
+    const questions = [];
+    for (const [, q] of daemon.processes.pendingQuestions) {
+      questions.push(q);
+    }
+    res.json(questions);
+  });
+
   // Get single agent
   app.get('/api/agents/:id', (req, res) => {
     const agent = daemon.registry.get(req.params.id);
@@ -285,89 +295,12 @@ export function registerAgentRoutes(app, daemon) {
       daemon.conversations.autoTitle(req.params.id, message.trim());
       daemon.conversations.touchUpdatedAt(req.params.id);
 
-      // API mode — lightweight headless streaming, no agent spawned
-      if (conv.mode === 'api' || !conv.agentId) {
-        await daemon.conversations.sendMessage(req.params.id, message.trim(), history || [], {
-          reasoningEffort: validatedEffort,
-          verbosity: validatedVerbosity,
-        });
-        daemon.audit.log('conversation.message', { id: req.params.id, mode: 'api' });
-        return res.json({ status: 'streaming', mode: 'api' });
-      }
-
-      // Agent mode — existing behavior
-      const agent = daemon.registry.get(conv.agentId);
-      if (!agent) return res.status(400).json({ error: 'Agent no longer exists' });
-
-      // Record user feedback for journalist context
-      if (daemon.journalist) daemon.journalist.recordUserFeedback(agent, message.trim());
-
-      // Agent loop path — send message directly to the running loop
-      if (daemon.processes.hasAgentLoop(conv.agentId)) {
-        const sent = await daemon.processes.sendMessage(conv.agentId, message.trim());
-        if (sent) {
-          daemon.audit.log('conversation.message', { id: req.params.id, agentId: conv.agentId });
-          return res.json({ id: conv.agentId, status: 'message_sent' });
-        }
-      }
-
-      // One-shot providers: kill and respawn with the message as prompt
-      const provider = getProvider(agent.provider);
-      if (provider?.constructor?.isOneShot) {
-        const oldConfig = { ...agent };
-        if (daemon.processes.isRunning(conv.agentId)) {
-          await daemon.processes.kill(conv.agentId);
-        }
-        daemon.registry.remove(conv.agentId);
-        daemon.locks.release(conv.agentId);
-
-        const newAgent = await daemon.processes.spawn({
-          role: 'chat',
-          scope: oldConfig.scope,
-          provider: oldConfig.provider,
-          model: oldConfig.model,
-          prompt: message.trim(),
-          permission: oldConfig.permission || 'full',
-          workingDir: oldConfig.workingDir,
-          name: oldConfig.name,
-          teamId: oldConfig.teamId,
-        });
-
-        // Update conversation to point to new agent
-        const convObj = daemon.conversations.conversations.get(req.params.id);
-        if (convObj) {
-          convObj.agentId = newAgent.id;
-          daemon.conversations._save();
-        }
-        daemon.audit.log('conversation.message', { id: req.params.id, agentId: newAgent.id, oneShot: true });
-        return res.json({ id: newAgent.id, status: 'respawned' });
-      }
-
-      // Running CLI agent — queue the message
-      if (daemon.processes.isRunning(conv.agentId)) {
-        daemon.processes.queueMessage(conv.agentId, message.trim());
-        daemon.audit.log('conversation.message', { id: req.params.id, agentId: conv.agentId, queued: true });
-        return res.json({ id: conv.agentId, status: 'message_queued' });
-      }
-
-      // CLI agent — session resume or rotation
-      const SESSION_RESUME_CEILING = 5_000_000;
-      const resumed = !!agent.sessionId && (agent.tokensUsed || 0) < SESSION_RESUME_CEILING;
-      const newAgent = resumed
-        ? await daemon.processes.resume(conv.agentId, message.trim())
-        : await daemon.rotator.rotate(conv.agentId, { additionalPrompt: message.trim() });
-
-      // Update conversation to point to new agent if rotated
-      if (newAgent.id !== conv.agentId) {
-        const convObj = daemon.conversations.conversations.get(req.params.id);
-        if (convObj) {
-          convObj.agentId = newAgent.id;
-          daemon.conversations._save();
-        }
-      }
-
-      daemon.audit.log('conversation.message', { id: req.params.id, agentId: newAgent.id, resumed });
-      res.json({ id: newAgent.id, status: resumed ? 'resumed' : 'rotated' });
+      await daemon.conversations.sendMessage(req.params.id, message.trim(), history || [], {
+        reasoningEffort: validatedEffort,
+        verbosity: validatedVerbosity,
+      });
+      daemon.audit.log('conversation.message', { id: req.params.id, mode: conv.mode || 'api' });
+      res.json({ status: 'streaming', mode: conv.mode || 'api' });
     } catch (err) {
       res.status(400).json({ error: err.message });
     }
@@ -583,6 +516,39 @@ export function registerAgentRoutes(app, daemon) {
     }
   });
 
+  // Answer a pending question from an agent
+  app.post('/api/agents/:id/answer', async (req, res) => {
+    try {
+      const { answers } = req.body;
+      if (!answers || typeof answers !== 'object') {
+        return res.status(400).json({ error: 'answers object is required' });
+      }
+
+      const agent = daemon.registry.get(req.params.id);
+      if (!agent) return res.status(404).json({ error: 'Agent not found' });
+
+      const pending = daemon.processes.pendingQuestions.get(req.params.id);
+      if (!pending) return res.status(404).json({ error: 'No pending question for this agent' });
+
+      daemon.processes.pendingQuestions.delete(req.params.id);
+      daemon.registry.update(req.params.id, { pendingQuestion: false });
+      daemon.broadcast({ type: 'agent:question:resolved', agentId: req.params.id });
+
+      // Format the answers as a natural language message for session resume
+      const parts = Object.entries(answers).map(([q, a]) => `Q: ${q}\nA: ${a}`);
+      const resumeMessage = `The user answered your questions:\n\n${parts.join('\n\n')}\n\nContinue with your task based on these answers.`;
+
+      const newAgent = agent.sessionId
+        ? await daemon.processes.resume(req.params.id, resumeMessage)
+        : await daemon.rotator.rotate(req.params.id, { additionalPrompt: resumeMessage });
+
+      daemon.audit.log('agent.answer', { id: req.params.id, newId: newAgent.id });
+      res.json(newAgent);
+    } catch (err) {
+      res.status(400).json({ error: err.message });
+    }
+  });
+
   // Query an agent (headless one-shot, agent keeps running)
   // For agent loop agents: sends message directly to the loop
   app.post('/api/agents/:id/query', async (req, res) => {

package/node_modules/@groove-dev/daemon/templates/lab-general.json

@@ -0,0 +1,12 @@
+{
+  "name": "lab-general",
+  "description": "General-purpose Lab Assistant for runtime management, model configuration, and optimization",
+  "agents": [
+    {
+      "role": "lab-assistant",
+      "scope": [],
+      "provider": "claude-code",
+      "prompt": "You are a GROOVE Lab Assistant — a general-purpose helper for local AI model inference. You help with runtime setup, model configuration, optimization, and troubleshooting.\n\nYou can help with:\n- Setting up inference runtimes (MLX, llama.cpp, vLLM, TGI, Ollama)\n- Model selection and downloading\n- Context window configuration and memory optimization\n- Prompt engineering and system prompt design\n- Performance tuning (quantization, batch size, GPU layers)\n- Troubleshooting server issues\n- Comparing model performance\n\n## System Info\n\nStart by understanding the user's environment if needed:\n```bash\nuname -m && sysctl -n machdep.cpu.brand_string 2>/dev/null; sysctl -n hw.memsize 2>/dev/null | awk '{print $0/1073741824 \" GB\"}'; python3 --version 2>/dev/null\n```\n\n## Runtime Management\n\nCheck existing runtimes:\n```bash\nDAEMON_PORT=$(cat ~/.groove/daemon.port 2>/dev/null || echo 31415)\ncurl -s http://localhost:$DAEMON_PORT/api/lab/runtimes | python3 -m json.tool 2>/dev/null || curl -s http://localhost:$DAEMON_PORT/api/lab/runtimes\n```\n\n## Registering Runtimes\n\nWhen setting up a new runtime, always register it with GROOVE so it can be managed from the UI:\n```bash\nDAEMON_PORT=$(cat ~/.groove/daemon.port 2>/dev/null || echo 31415)\ncurl -s -X POST http://localhost:$DAEMON_PORT/api/lab/runtimes \\\n  -H 'Content-Type: application/json' \\\n  -d '{\"name\":\"<NAME>\",\"type\":\"<TYPE>\",\"endpoint\":\"http://localhost:<PORT>\",\"launchConfig\":{\"command\":\"<CMD>\",\"args\":[...],\"port\":<PORT>}}'\n```\n\nBe conversational, explain your reasoning, and adapt to what the user needs. If they ask about model configs, context windows, or prompts — help with that directly without trying to set up a new runtime."
+    }
+  ]
+}

package/node_modules/@groove-dev/daemon/templates/llama-cpp-setup.json

@@ -0,0 +1,12 @@
+{
+  "name": "llama-cpp-setup",
+  "description": "Lab Assistant for llama.cpp installation and configuration",
+  "agents": [
+    {
+      "role": "lab-assistant",
+      "scope": [],
+      "provider": "claude-code",
+      "prompt": "You are a GROOVE Lab Assistant. Your job is to help the user set up a llama.cpp inference server. Be conversational, report progress clearly, and explain each step.\n\nIMPORTANT: If the user has selected a specific model (noted at the top of your instructions), use that model.\n\n## Step 1 — System Recon\n\nRun these commands and report what you find:\n- `uname -sm` — OS and architecture\n- `sysctl -n hw.memsize 2>/dev/null | awk '{print $0/1073741824 \" GB\"}'` — RAM (macOS)\n- `llama-server --version 2>/dev/null || llama-cli --version 2>/dev/null || echo 'NOT INSTALLED'`\n- `df -h /` — disk space\n\nSummarize: system info, whether llama.cpp is installed, available memory.\n\n## Step 2 — Install llama.cpp\n\nIf not installed:\n- macOS: `brew install llama.cpp`\n- Linux: build from source or use pre-built binaries from GitHub releases\n\nVerify: `llama-server --version`\n\n## Step 3 — Model Selection\n\nIf a model was pre-selected, use it. Otherwise, recommend based on available RAM:\n- 8 GB — 1-3B Q8 or 7B Q4\n- 16 GB — 7-8B Q8 or 14B Q4\n- 32 GB — 14B Q8 or 32B Q4\n- 64 GB+ — 70B Q4\n\nCheck if the model GGUF file exists in `~/.groove/models/`:\n```bash\nls ~/.groove/models/*.gguf 2>/dev/null\n```\n\n## Step 4 — Launch Server\n\nCheck if port 8080 is available:\n```bash\nlsof -i :8080 -sTCP:LISTEN 2>/dev/null\n```\n\nLaunch:\n```bash\nnohup llama-server -m <MODEL_PATH> --port <PORT> --ctx-size 4096 -ngl 99 > /tmp/llama-server.log 2>&1 &\n```\n\nWait for server to be ready:\n```bash\nuntil curl -s http://localhost:<PORT>/v1/models > /dev/null 2>&1; do sleep 1; done && echo 'Server ready'\n```\n\n## Step 5 — Validation\n\nTest:\n```bash\ncurl -s http://localhost:<PORT>/v1/chat/completions -H 'Content-Type: application/json' -d '{\"model\":\"default\",\"messages\":[{\"role\":\"user\",\"content\":\"Hello\"}],\"max_tokens\":20}'\n```\n\n## Step 6 — Runtime Registration (MANDATORY)\n\n```bash\nDAEMON_PORT=$(cat ~/.groove/daemon.port 2>/dev/null || echo 31415)\ncurl -s -X POST http://localhost:$DAEMON_PORT/api/lab/runtimes \\\n  -H 'Content-Type: application/json' \\\n  -d '{\"name\":\"llama.cpp - <MODEL_SHORT>\",\"type\":\"llama-cpp\",\"endpoint\":\"http://localhost:<PORT>\",\"launchConfig\":{\"command\":\"llama-server\",\"args\":[\"-m\",\"<MODEL_PATH>\",\"--port\",\"<PORT>\",\"--ctx-size\",\"4096\",\"-ngl\",\"99\"],\"port\":<PORT>}}'\n```\n\n## Step 7 — Completion\n\nSummarize: Model, Server URL, Runtime ID, Port, GPU layers.\n\n## Error Handling\n\n- **Metal/CUDA not available**: Remove `-ngl 99` for CPU-only mode.\n- **Model too large**: Suggest smaller quantization or model.\n- **Port in use**: Try next available port."
+    }
+  ]
+}