@stackmemoryai/stackmemory 0.5.30 → 0.5.33

package/README.md

@@ -1,19 +1,19 @@
 # StackMemory
 
-**Lossless, project-scoped memory for AI tools** • v0.5.29
+**Lossless, project-scoped memory for AI tools** • v0.5.30
 
 StackMemory is a **production-ready memory runtime** for AI coding tools that preserves full project context across sessions. With **Phases 1-4 complete**, it delivers:
 
-- ✅ **89-98% faster** task operations than manual tracking
-- ✅ **10,000+ frame depth** support with hierarchical organization
-- ✅ **Full Linear integration** with bidirectional sync
-- ✅ **20+ MCP tools** for Claude Code
-- ✅ **Context persistence** that survives /clear operations
-- ✅ **Two-tier storage system** with local tiers and infinite remote storage
-- ✅ **Smart compression** (LZ4/ZSTD) with 2.5-3.5x ratios
-- ✅ **Background migration** with configurable triggers
-- ✅ **396 tests passing** with standardized error handling
-- ✅ **npm v0.5.29** published with WhatsApp notifications and improved integrations
+- **89-98% faster** task operations than manual tracking
+- **10,000+ frame depth** support with hierarchical organization
+- **Full Linear integration** with bidirectional sync
+- **20+ MCP tools** for Claude Code
+- **Context persistence** that survives /clear operations
+- **Two-tier storage system** with local tiers and infinite remote storage
+- **Smart compression** (LZ4/ZSTD) with 2.5-3.5x ratios
+- **Background migration** with configurable triggers
+- **396 tests passing** with standardized error handling
+- **npm v0.5.30** published with WhatsApp notifications and improved integrations
 
 Instead of a linear chat log, StackMemory organizes memory as a **call stack** of scoped work (frames), with intelligent LLM-driven retrieval and team collaboration features.
 
@@ -96,16 +96,16 @@ The editor never manages memory directly; it asks StackMemory for the **context
 
 ## Product Health Metrics
 
-### Current Status (v0.5.29)
+### Current Status (v0.5.30)
 
 | Metric | Current | Target | Status |
 |--------|---------|--------|--------|
-| **Test Coverage** | 85% | 90% | 🟡 |
-| **Performance (p50)** | TBD | <50ms | 🔄 |
-| **Documentation** | 70% | 100% | 🟡 |
-| **Active Issues** | 5 high | 0 high | 🟡 |
-| **Code Quality** | 396 tests | 400+ | ✅ |
-| **npm Downloads** | Growing | 1K+/week | 🚀 |
+| **Test Coverage** | 85% | 90% | In Progress |
+| **Performance (p50)** | TBD | <50ms | Pending |
+| **Documentation** | 70% | 100% | In Progress |
+| **Active Issues** | 5 high | 0 high | In Progress |
+| **Code Quality** | 396 tests | 400+ | Done |
+| **npm Downloads** | Growing | 1K+/week | On Track |
 
 ### Quality Score: 78/100
 
@@ -136,7 +136,7 @@ This creates a **project-scoped memory space** tied to the repo.
 ### Step 2: Install StackMemory
 
 ```bash
-npm install -g @stackmemoryai/stackmemory@0.5.29
+npm install -g @stackmemoryai/stackmemory@0.5.30
 # or latest
 npm install -g @stackmemoryai/stackmemory@latest
 ```
@@ -455,7 +455,7 @@ stackmemory mcp-server [--port 3001]
 - Hosted: **Private beta**
 - OSS mirror: **Production ready**
 - MCP integration: **Stable**
-- CLI: **v0.5.29** - Full task, context, Linear, and storage management
+- CLI: **v0.5.30** - Full task, context, Linear, and storage management
 - Two-tier storage: **Complete**
 - Test Suite: **396 tests passing**
 
@@ -463,40 +463,40 @@ stackmemory mcp-server [--port 3001]
 
 ## Changelog
 
-### v0.5.29 (2026-01-26)
-- ✅ Standardized error handling with `IntegrationError`, `DatabaseError`, `ValidationError`
-- ✅ Adopted error classes across Linear integration (12 files)
-- ✅ Adopted error classes across database layer (6 files)
-- ✅ WhatsApp notifications with session ID and interactive options
-- ✅ 396 tests passing with improved code quality
+### v0.5.30 (2026-01-26)
+- Standardized error handling with `IntegrationError`, `DatabaseError`, `ValidationError`
+- Adopted error classes across Linear integration (12 files)
+- Adopted error classes across database layer (6 files)
+- WhatsApp notifications with session ID and interactive options
+- 396 tests passing with improved code quality
 
 ### v0.5.28 (2026-01-25)
-- ✅ WhatsApp flag for claude-sm automatic notifications
-- ✅ Incoming request queue for WhatsApp triggers
-- ✅ SMS webhook /send endpoint for outgoing notifications
+- WhatsApp flag for claude-sm automatic notifications
+- Incoming request queue for WhatsApp triggers
+- SMS webhook /send endpoint for outgoing notifications
 
 ### v0.5.26 (2026-01-24)
-- ✅ OpenCode wrapper (opencode-sm) with context integration
-- ✅ Discovery CLI and MCP tools
-- ✅ Real LLM provider and retrieval audit system
-- ✅ Linear issue management and task picker
+- OpenCode wrapper (opencode-sm) with context integration
+- Discovery CLI and MCP tools
+- Real LLM provider and retrieval audit system
+- Linear issue management and task picker
 
 ### v0.5.21 (2026-01-23)
-- ✅ Claude-sm remote mode and configurable defaults
-- ✅ Context loading command improvements
-- ✅ Session summary features
+- Claude-sm remote mode and configurable defaults
+- Context loading command improvements
+- Session summary features
 
 ### v0.3.16 (2026-01-15)
-- ✅ Fixed critical error handling - getFrame() returns undefined instead of throwing
-- ✅ Improved test coverage and fixed StackMemoryError constructor usage
-- ✅ Removed dangerous secret-cleaning scripts from repository
-- ✅ All tests passing, lint clean, build successful
+- Fixed critical error handling - getFrame() returns undefined instead of throwing
+- Improved test coverage and fixed StackMemoryError constructor usage
+- Removed dangerous secret-cleaning scripts from repository
+- All tests passing, lint clean, build successful
 
 ### v0.3.15 (2026-01-14)
-- ✅ Two-tier storage system implementation complete
-- ✅ Smart compression with LZ4/ZSTD support
-- ✅ Background migration with configurable triggers
-- ✅ Improved Linear integration with bidirectional sync
+- Two-tier storage system implementation complete
+- Smart compression with LZ4/ZSTD support
+- Background migration with configurable triggers
+- Improved Linear integration with bidirectional sync
 
 ---
 

package/dist/cli/claude-sm.js

@@ -17,7 +17,15 @@ import {
   generateSessionSummary,
   formatSummaryMessage
 } from "../hooks/session-summary.js";
-import { sendNotification } from "../hooks/sms-notify.js";
+import { sendNotification, loadSMSConfig } from "../hooks/sms-notify.js";
+import { enableAutoSync } from "../hooks/whatsapp-sync.js";
+import { enableCommands } from "../hooks/whatsapp-commands.js";
+import { startScheduler, listSchedules } from "../hooks/whatsapp-scheduler.js";
+import {
+  getModelRouter,
+  loadModelRouterConfig
+} from "../core/models/model-router.js";
+import { FallbackMonitor } from "../core/models/fallback-monitor.js";
 const DEFAULT_SM_CONFIG = {
   defaultWorktree: false,
   defaultSandbox: false,
@@ -25,7 +33,8 @@ const DEFAULT_SM_CONFIG = {
   defaultTracing: true,
   defaultRemote: false,
   defaultNotifyOnDone: true,
-  defaultWhatsApp: false
+  defaultWhatsApp: false,
+  defaultModelRouting: false
 };
 function getConfigPath() {
   return path.join(os.homedir(), ".stackmemory", "claude-sm.json");
@@ -68,7 +77,9 @@ class ClaudeSM {
       contextEnabled: true,
       tracingEnabled: this.smConfig.defaultTracing,
       verboseTracing: false,
-      sessionStartTime: Date.now()
+      sessionStartTime: Date.now(),
+      useModelRouting: this.smConfig.defaultModelRouting,
+      useThinkingMode: false
     };
     this.stackmemoryPath = this.findStackMemory();
     this.worktreeScriptPath = path.join(
@@ -249,7 +260,41 @@ class ClaudeSM {
   }
   async startWhatsAppServices() {
     const WEBHOOK_PORT = 3456;
-    console.log(chalk.cyan("Starting WhatsApp services..."));
+    console.log(chalk.cyan("\n\u{1F4F1} WhatsApp Mode - Full Integration"));
+    console.log(chalk.gray("\u2500".repeat(42)));
+    const smsConfig = loadSMSConfig();
+    if (!smsConfig.enabled) {
+      console.log(
+        chalk.yellow("  Notifications disabled. Run: stackmemory notify enable")
+      );
+    }
+    try {
+      enableAutoSync();
+      console.log(
+        chalk.green("  \u2713 Auto-sync enabled (digests on frame close)")
+      );
+    } catch {
+      console.log(chalk.gray("  Auto-sync: skipped"));
+    }
+    try {
+      enableCommands();
+      console.log(chalk.green("  \u2713 Command processing enabled"));
+      console.log(
+        chalk.gray("    Reply: status, tasks, context, help, build, test, lint")
+      );
+    } catch {
+      console.log(chalk.gray("  Command processing: skipped"));
+    }
+    try {
+      const schedules = listSchedules();
+      if (schedules.length > 0) {
+        startScheduler();
+        console.log(
+          chalk.green(`  \u2713 Scheduler started (${schedules.length} schedules)`)
+        );
+      }
+    } catch {
+    }
     const webhookRunning = await fetch(
       `http://localhost:${WEBHOOK_PORT}/health`
     ).then((r) => r.ok).catch(() => false);
@@ -262,11 +307,11 @@ class ClaudeSM {
       });
       webhookProcess.unref();
       console.log(
-        chalk.gray(`  Webhook server starting on port ${WEBHOOK_PORT}`)
+        chalk.green(`  \u2713 Webhook server started (port ${WEBHOOK_PORT})`)
       );
     } else {
       console.log(
-        chalk.gray(`  Webhook already running on port ${WEBHOOK_PORT}`)
+        chalk.green(`  \u2713 Webhook already running (port ${WEBHOOK_PORT})`)
       );
     }
     const ngrokRunning = await fetch("http://localhost:4040/api/tunnels").then((r) => r.ok).catch(() => false);
@@ -276,7 +321,7 @@ class ClaudeSM {
         stdio: "ignore"
       });
       ngrokProcess.unref();
-      console.log(chalk.gray("  ngrok tunnel starting..."));
+      console.log(chalk.gray("  Starting ngrok tunnel..."));
       await new Promise((resolve) => setTimeout(resolve, 3e3));
     }
     try {
@@ -291,15 +336,34 @@ class ClaudeSM {
           fs.mkdirSync(configDir, { recursive: true });
         }
         fs.writeFileSync(configPath, publicUrl);
-        console.log(
-          chalk.green(`  WhatsApp webhook: ${publicUrl}/sms/incoming`)
-        );
+        console.log(chalk.green(`  \u2713 ngrok tunnel: ${publicUrl}/sms/incoming`));
       }
     } catch {
+      console.log(chalk.yellow("  ngrok: waiting for tunnel..."));
+    }
+    if (smsConfig.pendingPrompts.length > 0) {
+      console.log();
       console.log(
-        chalk.yellow("  Waiting for ngrok... URL will be available shortly")
+        chalk.yellow(
+          `  \u23F3 ${smsConfig.pendingPrompts.length} pending prompt(s) awaiting response:`
+        )
       );
+      smsConfig.pendingPrompts.slice(0, 3).forEach((p, i) => {
+        const msg = p.message.length > 40 ? p.message.slice(0, 40) + "..." : p.message;
+        console.log(chalk.gray(`     ${i + 1}. [${p.id}] ${msg}`));
+      });
+      if (smsConfig.pendingPrompts.length > 3) {
+        console.log(
+          chalk.gray(`     ... and ${smsConfig.pendingPrompts.length - 3} more`)
+        );
+      }
     }
+    console.log();
+    console.log(chalk.gray("  Quick actions from WhatsApp:"));
+    console.log(chalk.gray('    "status"  - session status'));
+    console.log(chalk.gray('    "context" - current frame digest'));
+    console.log(chalk.gray('    "1", "2"  - respond to prompts'));
+    console.log(chalk.gray("\u2500".repeat(42)));
   }
   async sendDoneNotification(exitCode) {
     try {
@@ -430,6 +494,31 @@ class ClaudeSM {
             this.config.useWorktree = this.hasUncommittedChanges() || this.detectMultipleInstances();
           }
           break;
+        case "--think":
+        case "--think-hard":
+        case "--ultrathink":
+          this.config.useThinkingMode = true;
+          this.config.useModelRouting = true;
+          this.config.forceProvider = "qwen";
+          break;
+        case "--qwen":
+          this.config.useModelRouting = true;
+          this.config.forceProvider = "qwen";
+          break;
+        case "--openai":
+          this.config.useModelRouting = true;
+          this.config.forceProvider = "openai";
+          break;
+        case "--ollama":
+          this.config.useModelRouting = true;
+          this.config.forceProvider = "ollama";
+          break;
+        case "--model-routing":
+          this.config.useModelRouting = true;
+          break;
+        case "--no-model-routing":
+          this.config.useModelRouting = false;
+          break;
         default:
           claudeArgs.push(arg);
       }
@@ -517,6 +606,45 @@ class ClaudeSM {
         );
       }
     }
+    if (this.config.useModelRouting) {
+      const routerConfig = loadModelRouterConfig();
+      if (routerConfig.enabled || this.config.forceProvider) {
+        const router = getModelRouter();
+        let routeResult;
+        if (this.config.forceProvider) {
+          const env = router.switchTo(this.config.forceProvider);
+          Object.assign(process.env, env);
+          console.log(
+            chalk.magenta(`\u{1F500} Model: ${this.config.forceProvider} (forced)`)
+          );
+          if (this.config.forceProvider === "qwen" && this.config.useThinkingMode) {
+            const qwenConfig = routerConfig.providers.qwen;
+            if (qwenConfig?.params?.enable_thinking) {
+              console.log(
+                chalk.gray(
+                  `   Thinking mode: budget ${qwenConfig.params.thinking_budget || 1e4} tokens`
+                )
+              );
+            }
+          }
+        } else {
+          const taskType = this.config.useThinkingMode ? "think" : "default";
+          routeResult = router.route(taskType, this.config.task);
+          Object.assign(process.env, routeResult.env);
+          if (routeResult.switched) {
+            console.log(
+              chalk.magenta(`\u{1F500} Model routed to: ${routeResult.provider}`)
+            );
+          }
+        }
+      } else {
+        console.log(
+          chalk.gray(
+            "   Model routing: disabled (run: stackmemory model enable)"
+          )
+        );
+      }
+    }
     if (this.config.useWhatsApp) {
       console.log(
         chalk.cyan("\u{1F4F1} WhatsApp mode: notifications + webhook enabled")
@@ -537,10 +665,36 @@ class ClaudeSM {
       process.exit(1);
       return;
     }
-    const claude = spawn(claudeBin, claudeArgs, {
-      stdio: "inherit",
-      env: process.env
+    const fallbackMonitor = new FallbackMonitor({
+      enabled: true,
+      maxRestarts: 2,
+      restartDelayMs: 1500,
+      onFallback: (provider, reason) => {
+        console.log(chalk.yellow(`
+[auto-fallback] Switching to ${provider}`));
+        console.log(chalk.gray(`   Reason: ${reason}`));
+        console.log(chalk.gray(`   Session will continue on ${provider}...`));
+        if (this.config.notifyOnDone || this.config.useWhatsApp) {
+          sendNotification({
+            type: "custom",
+            title: "Model Fallback",
+            message: `Claude unavailable (${reason}). Switched to ${provider}.`
+          }).catch(() => {
+          });
+        }
+      }
     });
+    const fallbackAvailable = fallbackMonitor.isFallbackAvailable();
+    if (fallbackAvailable) {
+      console.log(
+        chalk.gray(`   Auto-fallback: Qwen ready (on rate limit/error)`)
+      );
+    }
+    const wrapper = fallbackMonitor.wrapProcess(claudeBin, claudeArgs, {
+      env: process.env,
+      cwd: process.cwd()
+    });
+    const claude = wrapper.start();
     claude.on("error", (err) => {
       console.error(chalk.red("\u274C Failed to launch Claude CLI."));
       if (err.code === "ENOENT") {
@@ -559,6 +713,15 @@ class ClaudeSM {
       process.exit(1);
     });
     claude.on("exit", async (code) => {
+      const status = fallbackMonitor.getStatus();
+      if (status.inFallback) {
+        console.log(
+          chalk.yellow(
+            `
+Session completed on fallback provider: ${status.currentProvider}`
+          )
+        );
+      }
       this.saveContext("Claude session ended", {
         action: "session_end",
         exitCode: code
@@ -624,6 +787,9 @@ configCmd.command("show").description("Show current default settings").action(()
   console.log(
     `  defaultWhatsApp: ${config.defaultWhatsApp ? chalk.green("true") : chalk.gray("false")}`
   );
+  console.log(
+    `  defaultModelRouting: ${config.defaultModelRouting ? chalk.green("true") : chalk.gray("false")}`
+  );
   console.log(chalk.gray(`
 Config: ${getConfigPath()}`));
 });
@@ -638,7 +804,9 @@ configCmd.command("set <key> <value>").description("Set a default (e.g., set wor
     remote: "defaultRemote",
     "notify-done": "defaultNotifyOnDone",
     notifyondone: "defaultNotifyOnDone",
-    whatsapp: "defaultWhatsApp"
+    whatsapp: "defaultWhatsApp",
+    "model-routing": "defaultModelRouting",
+    modelrouting: "defaultModelRouting"
   };
   const configKey = keyMap[key];
   if (!configKey) {
@@ -704,10 +872,25 @@ configCmd.command("whatsapp-off").description("Disable WhatsApp mode by default"
   saveSMConfig(config);
   console.log(chalk.green("WhatsApp mode disabled by default"));
 });
+configCmd.command("model-routing-on").description(
+  "Enable model routing by default (route tasks to Qwen/other models)"
+).action(() => {
+  const config = loadSMConfig();
+  config.defaultModelRouting = true;
+  saveSMConfig(config);
+  console.log(chalk.green("Model routing enabled by default"));
+  console.log(chalk.gray("Configure with: stackmemory model setup-qwen"));
+});
+configCmd.command("model-routing-off").description("Disable model routing by default (use Claude only)").action(() => {
+  const config = loadSMConfig();
+  config.defaultModelRouting = false;
+  saveSMConfig(config);
+  console.log(chalk.green("Model routing disabled by default"));
+});
 program.option("-w, --worktree", "Create isolated worktree for this instance").option("-W, --no-worktree", "Disable worktree (override default)").option("-r, --remote", "Enable remote mode (WhatsApp for all questions)").option("--no-remote", "Disable remote mode (override default)").option("-n, --notify-done", "Send WhatsApp notification when session ends").option("--no-notify-done", "Disable notification when session ends").option(
   "--whatsapp",
   "Enable WhatsApp mode (auto-start webhook + ngrok + notifications)"
-).option("--no-whatsapp", "Disable WhatsApp mode (override default)").option("-s, --sandbox", "Enable sandbox mode (file/network restrictions)").option("-c, --chrome", "Enable Chrome automation").option("-a, --auto", "Automatically detect and apply best settings").option("-b, --branch <name>", "Specify branch name for worktree").option("-t, --task <desc>", "Task description for context").option("--claude-bin <path>", "Path to claude CLI (or use CLAUDE_BIN)").option("--no-context", "Disable StackMemory context integration").option("--no-trace", "Disable debug tracing (enabled by default)").option("--verbose-trace", "Enable verbose debug tracing with full details").helpOption("-h, --help", "Display help").allowUnknownOption(true).action(async (_options) => {
+).option("--no-whatsapp", "Disable WhatsApp mode (override default)").option("-s, --sandbox", "Enable sandbox mode (file/network restrictions)").option("-c, --chrome", "Enable Chrome automation").option("-a, --auto", "Automatically detect and apply best settings").option("-b, --branch <name>", "Specify branch name for worktree").option("-t, --task <desc>", "Task description for context").option("--claude-bin <path>", "Path to claude CLI (or use CLAUDE_BIN)").option("--no-context", "Disable StackMemory context integration").option("--no-trace", "Disable debug tracing (enabled by default)").option("--verbose-trace", "Enable verbose debug tracing with full details").option("--think", "Enable thinking mode with Qwen (deep reasoning)").option("--think-hard", "Alias for --think").option("--ultrathink", "Alias for --think").option("--qwen", "Force Qwen provider for this session").option("--openai", "Force OpenAI provider for this session").option("--ollama", "Force Ollama provider for this session").option("--model-routing", "Enable model routing").option("--no-model-routing", "Disable model routing").helpOption("-h, --help", "Display help").allowUnknownOption(true).action(async (_options) => {
   const claudeSM = new ClaudeSM();
   const args = process.argv.slice(2);
   await claudeSM.run(args);