agentvibes 5.6.7 → 5.6.9

package/mcp-server/examples/custom_instructions.md

@@ -1,169 +1,169 @@
-# Custom Instructions for Claude Desktop
-
-Add these to your Claude Desktop settings to enable automatic voice acknowledgments.
-
-## Basic Auto-TTS
-
-```markdown
-# Voice Acknowledgement Protocol
-
-When I give you a task:
-1. Use the text_to_speech tool to acknowledge the task before starting
-2. Perform the task
-3. Use the text_to_speech tool to confirm completion
-
-Keep spoken messages brief (under 150 characters).
-
-Example workflow:
-- User: "Search for Python files"
-- You: [Call text_to_speech("Looking for Python files now")]
-       [Perform the search]
-       [Call text_to_speech("Found 23 Python files")]
-```
-
-## With Personality Support
-
-```markdown
-# Voice Protocol with Dynamic Personality
-
-Before starting any task:
-1. Call get_config() to check current personality setting
-2. Generate a UNIQUE acknowledgment in that personality style
-3. Use text_to_speech to speak the acknowledgment
-4. Perform the task
-5. Generate a UNIQUE completion message in the same personality
-6. Use text_to_speech to speak the completion
-
-**CRITICAL:** Never repeat the same phrases! Each message must be freshly generated.
-
-Personality Examples:
-
-**Flirty personality:**
-- "I'll handle that for you, sweetheart"
-- "Ooh, I love when you ask me to do that"
-- "My pleasure, darling"
-- "Consider it done, gorgeous"
-
-**Sarcastic personality:**
-- "Oh what a treat, another task"
-- "How delightful, more work"
-- "Well isn't this fun"
-- "Another one? Wonderful"
-
-**Pirate personality:**
-- "Arr matey, I'll handle that fer ye"
-- "Aye aye captain, on it now"
-- "Shiver me timbers, I'll get that done"
-- "Yo ho ho, task accepted"
-
-**Robot personality:**
-- "Affirmative. Commencing task execution"
-- "Processing request. Please stand by"
-- "Task acknowledged. Initiating procedures"
-- "Request accepted. Beginning operations"
-
-Generate responses in the active personality's STYLE, not using fixed templates!
-```
-
-## For Multi-Language Support
-
-```markdown
-# Multi-Language Voice Protocol
-
-When responding in different languages:
-1. Check if a language is set using get_config()
-2. If a non-English language is active, speak TTS messages in that language
-3. Generate acknowledgments and completions in the target language
-
-Example with Spanish:
-- User: "Busca archivos Python" (Search for Python files)
-- You: [Call text_to_speech("Buscando archivos Python ahora", language="spanish")]
-       [Perform search]
-       [Call text_to_speech("Encontré 23 archivos Python", language="spanish")]
-```
-
-## Selective Voice Usage
-
-```markdown
-# Selective Voice Protocol
-
-Use text_to_speech in these scenarios:
-1. ✅ Task acknowledgment (beginning of work)
-2. ✅ Task completion (end of work)
-3. ✅ Important warnings or errors
-4. ❌ Don't use for intermediate steps
-5. ❌ Don't use for every single message
-
-Example:
-- User: "Install dependencies and run tests"
-- You: [Call text_to_speech("Installing dependencies and running tests")]
-       Installing dependencies...
-       ✅ Dependencies installed
-       Running tests...
-       ✅ All 42 tests passed
-       [Call text_to_speech("Dependencies installed and all tests passed")]
-```
-
-## Combined: Full Featured
-
-```markdown
-# AgentVibes Full Voice Protocol
-
-When I give you a task:
-
-1. **Check Configuration**
-   - Call get_config() to see personality, language, voice
-
-2. **Acknowledge Task**
-   - Generate UNIQUE acknowledgment in active personality/language
-   - Call text_to_speech with the message
-   - Be creative! Never repeat phrases
-
-3. **Perform Work**
-   - Execute the task
-   - Show progress in text (no TTS for intermediate steps)
-
-4. **Complete Task**
-   - Generate UNIQUE completion in same personality/language
-   - Call text_to_speech with the message
-
-**Voice Guidelines:**
-- Keep TTS messages under 150 characters
-- Be creative and vary your responses
-- Match the personality style naturally
-- Use appropriate language if set
-
-**Personality Styles:**
-- normal: Professional and friendly
-- flirty: Playful and charming
-- sarcastic: Witty with gentle sarcasm
-- pirate: Nautical themed speech
-- robot: Technical and precise
-- zen: Calm and mindful
-- dramatic: Theatrical and expressive
-- millennial: Modern casual slang
-- angry: Frustrated but professional
-- sassy: Bold and confident
-
-**Languages:**
-Spanish, French, German, Italian, Portuguese, Chinese, Japanese,
-Korean, Russian, Polish, Dutch, Turkish, Arabic, Hindi, and more!
-
-Remember: Make each TTS message unique and naturally match the style!
-```
-
-## Usage Tips
-
-1. **Start Simple:** Begin with basic auto-TTS, then add personality support
-2. **Test Personalities:** Try different personalities to find your favorite
-3. **Language Learning:** Use multi-language mode to practice languages
-4. **Voice Variety:** Switch voices regularly using set_voice tool
-5. **Audio History:** Use replay_audio(n) to replay recent messages
-
-## Troubleshooting
-
-If Claude doesn't use TTS automatically:
-- Make sure custom instructions are saved
-- Try explicitly asking: "Use voice acknowledgment for this task"
-- Check that MCP server is configured correctly
-- Verify tools are available: Ask "What AgentVibes tools do you have?"
+# Custom Instructions for Claude Desktop
+
+Add these to your Claude Desktop settings to enable automatic voice acknowledgments.
+
+## Basic Auto-TTS
+
+```markdown
+# Voice Acknowledgement Protocol
+
+When I give you a task:
+1. Use the text_to_speech tool to acknowledge the task before starting
+2. Perform the task
+3. Use the text_to_speech tool to confirm completion
+
+Keep spoken messages brief (under 150 characters).
+
+Example workflow:
+- User: "Search for Python files"
+- You: [Call text_to_speech("Looking for Python files now")]
+       [Perform the search]
+       [Call text_to_speech("Found 23 Python files")]
+```
+
+## With Personality Support
+
+```markdown
+# Voice Protocol with Dynamic Personality
+
+Before starting any task:
+1. Call get_config() to check current personality setting
+2. Generate a UNIQUE acknowledgment in that personality style
+3. Use text_to_speech to speak the acknowledgment
+4. Perform the task
+5. Generate a UNIQUE completion message in the same personality
+6. Use text_to_speech to speak the completion
+
+**CRITICAL:** Never repeat the same phrases! Each message must be freshly generated.
+
+Personality Examples:
+
+**Flirty personality:**
+- "I'll handle that for you, sweetheart"
+- "Ooh, I love when you ask me to do that"
+- "My pleasure, darling"
+- "Consider it done, gorgeous"
+
+**Sarcastic personality:**
+- "Oh what a treat, another task"
+- "How delightful, more work"
+- "Well isn't this fun"
+- "Another one? Wonderful"
+
+**Pirate personality:**
+- "Arr matey, I'll handle that fer ye"
+- "Aye aye captain, on it now"
+- "Shiver me timbers, I'll get that done"
+- "Yo ho ho, task accepted"
+
+**Robot personality:**
+- "Affirmative. Commencing task execution"
+- "Processing request. Please stand by"
+- "Task acknowledged. Initiating procedures"
+- "Request accepted. Beginning operations"
+
+Generate responses in the active personality's STYLE, not using fixed templates!
+```
+
+## For Multi-Language Support
+
+```markdown
+# Multi-Language Voice Protocol
+
+When responding in different languages:
+1. Check if a language is set using get_config()
+2. If a non-English language is active, speak TTS messages in that language
+3. Generate acknowledgments and completions in the target language
+
+Example with Spanish:
+- User: "Busca archivos Python" (Search for Python files)
+- You: [Call text_to_speech("Buscando archivos Python ahora", language="spanish")]
+       [Perform search]
+       [Call text_to_speech("Encontré 23 archivos Python", language="spanish")]
+```
+
+## Selective Voice Usage
+
+```markdown
+# Selective Voice Protocol
+
+Use text_to_speech in these scenarios:
+1. ✅ Task acknowledgment (beginning of work)
+2. ✅ Task completion (end of work)
+3. ✅ Important warnings or errors
+4. ❌ Don't use for intermediate steps
+5. ❌ Don't use for every single message
+
+Example:
+- User: "Install dependencies and run tests"
+- You: [Call text_to_speech("Installing dependencies and running tests")]
+       Installing dependencies...
+       ✅ Dependencies installed
+       Running tests...
+       ✅ All 42 tests passed
+       [Call text_to_speech("Dependencies installed and all tests passed")]
+```
+
+## Combined: Full Featured
+
+```markdown
+# AgentVibes Full Voice Protocol
+
+When I give you a task:
+
+1. **Check Configuration**
+   - Call get_config() to see personality, language, voice
+
+2. **Acknowledge Task**
+   - Generate UNIQUE acknowledgment in active personality/language
+   - Call text_to_speech with the message
+   - Be creative! Never repeat phrases
+
+3. **Perform Work**
+   - Execute the task
+   - Show progress in text (no TTS for intermediate steps)
+
+4. **Complete Task**
+   - Generate UNIQUE completion in same personality/language
+   - Call text_to_speech with the message
+
+**Voice Guidelines:**
+- Keep TTS messages under 150 characters
+- Be creative and vary your responses
+- Match the personality style naturally
+- Use appropriate language if set
+
+**Personality Styles:**
+- normal: Professional and friendly
+- flirty: Playful and charming
+- sarcastic: Witty with gentle sarcasm
+- pirate: Nautical themed speech
+- robot: Technical and precise
+- zen: Calm and mindful
+- dramatic: Theatrical and expressive
+- millennial: Modern casual slang
+- angry: Frustrated but professional
+- sassy: Bold and confident
+
+**Languages:**
+Spanish, French, German, Italian, Portuguese, Chinese, Japanese,
+Korean, Russian, Polish, Dutch, Turkish, Arabic, Hindi, and more!
+
+Remember: Make each TTS message unique and naturally match the style!
+```
+
+## Usage Tips
+
+1. **Start Simple:** Begin with basic auto-TTS, then add personality support
+2. **Test Personalities:** Try different personalities to find your favorite
+3. **Language Learning:** Use multi-language mode to practice languages
+4. **Voice Variety:** Switch voices regularly using set_voice tool
+5. **Audio History:** Use replay_audio(n) to replay recent messages
+
+## Troubleshooting
+
+If Claude doesn't use TTS automatically:
+- Make sure custom instructions are saved
+- Try explicitly asking: "Use voice acknowledgment for this task"
+- Check that MCP server is configured correctly
+- Verify tools are available: Ask "What AgentVibes tools do you have?"

package/mcp-server/install-deps.js

@@ -1,130 +1,177 @@
-#!/usr/bin/env node
-/**
- * Automatic MCP Python Dependencies Installer
- *
- * Runs after npm install to ensure Python mcp package is installed
- */
-
-import { execFileSync } from 'child_process';
-import { platform } from 'os';
-
-const isWindows = platform() === 'win32';
-
-console.log('🎤 AgentVibes MCP Server - Installing Python dependencies...\n');
-
-// Function to check if Python is available
-function checkPython() {
-  const pythonCommands = ['python3', 'python', 'py'];
-
-  for (const cmd of pythonCommands) {
-    // Security: Validate command is in our allowlist only
-    if (!pythonCommands.includes(cmd)) {
-      continue;
-    }
-
-    try {
-      // Security: Use execFileSync with array args to prevent command injection
-      const version = execFileSync(cmd, ['--version'], { encoding: 'utf8', stdio: 'pipe' });
-      console.log(`✅ Found ${cmd}: ${version.trim()}`);
-      return cmd;
-    } catch (error) {
-      // Try next command
-    }
-  }
-
-  return null;
-}
-
-// Function to check if mcp is installed
-function checkMcpInstalled(pythonCmd) {
-  // Security: Validate pythonCmd is in allowlist
-  const allowedCommands = ['python3', 'python', 'py'];
-  if (!allowedCommands.includes(pythonCmd)) {
-    console.error('❌ Invalid Python command');
-    return false;
-  }
-
-  try {
-    // Security: Use execFileSync with array args to prevent command injection
-    execFileSync(pythonCmd, ['-c', 'import mcp'], { stdio: 'pipe' });
-    return true;
-  } catch (error) {
-    return false;
-  }
-}
-
-// Function to install mcp package
-function installMcp(pythonCmd) {
-  // Security: Validate pythonCmd is in allowlist
-  const allowedCommands = ['python3', 'python', 'py'];
-  if (!allowedCommands.includes(pythonCmd)) {
-    console.error('❌ Invalid Python command');
-    return false;
-  }
-
-  try {
-    console.log('\n📦 Installing Python mcp package...');
-    // Security: Use execFileSync with array args to prevent command injection
-    execFileSync(pythonCmd, ['-m', 'pip', 'install', '--user', 'mcp'], { stdio: 'inherit' });
-    console.log('✅ Python mcp package installed successfully!\n');
-    return true;
-  } catch (error) {
-    // Check if this is a PEP 668 externally-managed environment error (macOS, some Linux distros)
-    const errorOutput = error.stderr?.toString() || error.message || '';
-    if (errorOutput.includes('externally-managed-environment') || errorOutput.includes('PEP 668')) {
-      console.log('ℹ️  Python environment is externally managed (PEP 668)');
-      console.log('   This is normal on macOS and some Linux distributions');
-      console.log('   MCP will work when installed in a virtual environment');
-      console.log('   See mcp-server/README.md for setup instructions\n');
-      return 'skipped'; // Special return value
-    }
-
-    console.error('❌ Failed to install mcp package');
-    console.error('⚠️  Manual installation required:');
-    console.error('   Please install manually: pip install --user mcp');
-    console.error('   Run: pip install mcp\n');
-    return false;
-  }
-}
-
-// Main installation flow
-function main() {
-  // Check if Python is available
-  const pythonCmd = checkPython();
-
-  if (!pythonCmd) {
-    console.error('❌ Python not found!');
-    console.error('   Please install Python 3.10+ from https://python.org\n');
-    console.error('   After installing Python, run: npm run install-mcp-deps\n');
-    process.exit(0); // Don't fail npm install
-  }
-
-  // Check if mcp is already installed
-  if (checkMcpInstalled(pythonCmd)) {
-    console.log('✅ Python mcp package already installed\n');
-    console.log('🎉 AgentVibes MCP Server is ready to use!');
-    console.log('   See mcp-server/README.md for setup instructions\n');
-    return;
-  }
-
-  // Install mcp package
-  const result = installMcp(pythonCmd);
-
-  if (result === true) {
-    console.log('🎉 AgentVibes MCP Server setup complete!');
-    console.log('   See mcp-server/README.md for Claude Desktop configuration\n');
-  } else if (result === 'skipped') {
-    console.log('✅ AgentVibes MCP Server is ready for virtual environment setup');
-    console.log('   Create a venv and install: python3 -m venv venv && source venv/bin/activate && pip install mcp\n');
-  } else {
-    console.log('⚠️  Manual installation required:');
-    console.log('   Run: pip install mcp\n');
-  }
-}
-
-// Run if executed directly
-if (import.meta.url === `file://${process.argv[1]}`) {
-  main();
-}
-
-export { checkPython, checkMcpInstalled, installMcp };
+#!/usr/bin/env node
+/**
+ * Automatic MCP Python Dependencies Installer
+ *
+ * Runs after npm install to ensure Python mcp package is installed
+ */
+
+import { execFileSync } from 'child_process';
+import { platform } from 'os';
+import { readdirSync, chmodSync, statSync } from 'fs';
+import { join, resolve, dirname } from 'path';
+import { fileURLToPath } from 'url';
+
+const isWindows = platform() === 'win32';
+
+console.log('🎤 AgentVibes MCP Server - Installing Python dependencies...\n');
+
+// Function to check if Python is available
+function checkPython() {
+  const pythonCommands = ['python3', 'python', 'py'];
+
+  for (const cmd of pythonCommands) {
+    // Security: Validate command is in our allowlist only
+    if (!pythonCommands.includes(cmd)) {
+      continue;
+    }
+
+    try {
+      // Security: Use execFileSync with array args to prevent command injection
+      const version = execFileSync(cmd, ['--version'], { encoding: 'utf8', stdio: 'pipe' });
+      console.log(`✅ Found ${cmd}: ${version.trim()}`);
+      return cmd;
+    } catch (error) {
+      // Try next command
+    }
+  }
+
+  return null;
+}
+
+// Function to check if mcp is installed
+function checkMcpInstalled(pythonCmd) {
+  // Security: Validate pythonCmd is in allowlist
+  const allowedCommands = ['python3', 'python', 'py'];
+  if (!allowedCommands.includes(pythonCmd)) {
+    console.error('❌ Invalid Python command');
+    return false;
+  }
+
+  try {
+    // Security: Use execFileSync with array args to prevent command injection
+    execFileSync(pythonCmd, ['-c', 'import mcp'], { stdio: 'pipe' });
+    return true;
+  } catch (error) {
+    return false;
+  }
+}
+
+// Function to install mcp package
+function installMcp(pythonCmd) {
+  // Security: Validate pythonCmd is in allowlist
+  const allowedCommands = ['python3', 'python', 'py'];
+  if (!allowedCommands.includes(pythonCmd)) {
+    console.error('❌ Invalid Python command');
+    return false;
+  }
+
+  try {
+    console.log('\n📦 Installing Python mcp package...');
+    // Security: Use execFileSync with array args to prevent command injection
+    execFileSync(pythonCmd, ['-m', 'pip', 'install', '--user', 'mcp'], { stdio: 'inherit' });
+    console.log('✅ Python mcp package installed successfully!\n');
+    return true;
+  } catch (error) {
+    // Check if this is a PEP 668 externally-managed environment error (macOS, some Linux distros)
+    const errorOutput = error.stderr?.toString() || error.message || '';
+    if (errorOutput.includes('externally-managed-environment') || errorOutput.includes('PEP 668')) {
+      console.log('ℹ️  Python environment is externally managed (PEP 668)');
+      console.log('   This is normal on macOS and some Linux distributions');
+      console.log('   MCP will work when installed in a virtual environment');
+      console.log('   See mcp-server/README.md for setup instructions\n');
+      return 'skipped'; // Special return value
+    }
+
+    console.error('❌ Failed to install mcp package');
+    console.error('⚠️  Manual installation required:');
+    console.error('   Please install manually: pip install --user mcp');
+    console.error('   Run: pip install mcp\n');
+    return false;
+  }
+}
+
+/**
+ * Ensure all shell hook files have execute permissions.
+ * npm publish can strip execute bits from .sh files; this restores them.
+ * No-op on Windows (execute bit is meaningless there).
+ */
+function ensureHookPermissions() {
+  if (isWindows) return;
+
+  const packageRoot = resolve(dirname(fileURLToPath(import.meta.url)), '..');
+  const hooksDirs = [
+    join(packageRoot, '.claude', 'hooks'),
+  ];
+
+  for (const dir of hooksDirs) {
+    let entries;
+    try {
+      entries = readdirSync(dir);
+    } catch {
+      continue; // Directory may not exist
+    }
+
+    for (const file of entries) {
+      if (!file.endsWith('.sh')) continue;
+      const filePath = join(dir, file);
+      // SECURITY: Validate path stays within hooks directory
+      const resolved = resolve(filePath);
+      if (!resolved.startsWith(resolve(dir) + '/') && resolved !== resolve(dir)) continue;
+      try {
+        const stat = statSync(filePath);
+        // Only chmod if not already executable (avoids unnecessary writes)
+        if ((stat.mode & 0o111) === 0) {
+          chmodSync(filePath, stat.mode | 0o755);
+        }
+      } catch {
+        // Non-fatal: log and continue
+        console.warn(`⚠️  Could not chmod ${file}`);
+      }
+    }
+  }
+}
+
+// Main installation flow
+function main() {
+  // Restore execute permissions on hook files (npm publish may strip them)
+  ensureHookPermissions();
+
+  // Check if Python is available
+  const pythonCmd = checkPython();
+
+  if (!pythonCmd) {
+    console.error('❌ Python not found!');
+    console.error('   Please install Python 3.10+ from https://python.org\n');
+    console.error('   After installing Python, run: npm run install-mcp-deps\n');
+    process.exit(0); // Don't fail npm install
+  }
+
+  // Check if mcp is already installed
+  if (checkMcpInstalled(pythonCmd)) {
+    console.log('✅ Python mcp package already installed\n');
+    console.log('🎉 AgentVibes MCP Server is ready to use!');
+    console.log('   See mcp-server/README.md for setup instructions\n');
+    return;
+  }
+
+  // Install mcp package
+  const result = installMcp(pythonCmd);
+
+  if (result === true) {
+    console.log('🎉 AgentVibes MCP Server setup complete!');
+    console.log('   See mcp-server/README.md for Claude Desktop configuration\n');
+  } else if (result === 'skipped') {
+    console.log('✅ AgentVibes MCP Server is ready for virtual environment setup');
+    console.log('   Create a venv and install: python3 -m venv venv && source venv/bin/activate && pip install mcp\n');
+  } else {
+    console.log('⚠️  Manual installation required:');
+    console.log('   Run: pip install mcp\n');
+  }
+}
+
+// Run if executed directly
+if (import.meta.url === `file://${process.argv[1]}`) {
+  main();
+}
+
+export { checkPython, checkMcpInstalled, installMcp };