@mndrk/agx 1.0.0

package/README.md

@@ -0,0 +1,83 @@
+# agx
+
+Unified AI Agent Wrapper for Gemini, Claude, and Ollama.
+
+## Installation
+
+From the `agx` directory:
+```bash
+npm link
+```
+Now you can use `agx` globally.
+
+## Usage
+
+```bash
+agx <provider> [options] --prompt "<prompt>"
+```
+
+### Providers
+
+| Provider | Aliases | Backend |
+|----------|---------|---------|
+| `gemini` | `gem`, `g` | Google Gemini CLI |
+| `claude` | `cl`, `c` | Anthropic Claude CLI |
+| `ollama` | `ol`, `o` | Local Ollama via Claude interface |
+
+### Options
+
+| Option | Short | Description |
+|--------|-------|-------------|
+| `--prompt <text>` | `-p` | The prompt to send |
+| `--model <name>` | `-m` | Model name to use |
+| `--yolo` | `-y` | Skip permission prompts |
+| `--print` | | Non-interactive mode (output and exit) |
+| `--interactive` | `-i` | Force interactive mode |
+| `--sandbox` | `-s` | Enable sandbox (gemini only) |
+| `--debug` | `-d` | Enable debug output |
+| `--mcp <config>` | | MCP config file (claude/ollama only) |
+
+### Raw Passthrough
+
+Use `--` to pass arguments directly to the underlying CLI:
+```bash
+agx claude -- --resume
+```
+
+## LLM-Predictable Command Patterns
+
+For LLMs constructing commands, use these canonical patterns:
+
+```bash
+# Pattern: agx <provider> --prompt "<prompt>"
+agx claude --prompt "explain this code"
+agx gemini --prompt "summarize the file"
+agx ollama --prompt "write a function"
+
+# Pattern: agx <provider> --model <model> --prompt "<prompt>"
+agx claude --model claude-sonnet-4-20250514 --prompt "fix the bug"
+agx gemini --model gemini-2.0-flash --prompt "optimize this"
+agx ollama --model qwen3:8b --prompt "refactor"
+
+# Pattern: agx <provider> --yolo --prompt "<prompt>"
+agx claude --yolo --prompt "run the tests"
+
+# Pattern: agx <provider> --print --prompt "<prompt>"
+agx claude --print --prompt "what is 2+2"
+```
+
+### Command Structure
+
+```
+agx <provider> [--model <name>] [--yolo] [--print] --prompt "<prompt>"
+```
+
+**Rules for LLMs:**
+1. Always use `--prompt` flag for the prompt text
+2. Quote the prompt with double quotes
+3. Place options before `--prompt`
+4. Use full provider names (`claude`, `gemini`, `ollama`) for clarity
+
+## Ollama Support
+
+`agx ollama` automatically configures the environment to use a local Ollama instance as the backend for Claude Code. Default model is `glm-4.7:cloud` unless specified with `--model`.

package/index.js

@@ -0,0 +1,228 @@
+#!/usr/bin/env node
+
+const { spawn } = require('child_process');
+
+const args = process.argv.slice(2);
+let provider = args[0];
+
+// Normalize provider aliases
+const PROVIDER_ALIASES = {
+  'g': 'gemini',
+  'gem': 'gemini',
+  'gemini': 'gemini',
+  'c': 'claude',
+  'cl': 'claude',
+  'claude': 'claude',
+  'o': 'ollama',
+  'ol': 'ollama',
+  'ollama': 'ollama'
+};
+
+const VALID_PROVIDERS = ['gemini', 'claude', 'ollama'];
+
+// Handle help/version before provider check
+if (args.includes('--help') || args.includes('-h') || !provider) {
+  console.log(`agx - Unified AI Agent CLI
+
+SYNTAX:
+  agx <provider> [options] "<prompt>"
+  agx <provider> [options] --prompt "<prompt>"
+
+PROVIDERS:
+  gemini, gem, g     Google Gemini
+  claude, cl, c      Anthropic Claude
+  ollama, ol, o      Local Ollama (via Claude interface)
+
+OPTIONS:
+  --prompt, -p <text>    The prompt to send (recommended for clarity)
+  --model, -m <name>     Model name to use
+  --yolo, -y             Skip permission prompts
+  --print                Non-interactive mode (output and exit)
+  --interactive, -i      Force interactive mode
+  --sandbox, -s          Enable sandbox (gemini only)
+  --debug, -d            Enable debug output
+  --mcp <config>         MCP config file (claude/ollama only)
+
+EXAMPLES:
+  agx claude --prompt "explain this code"
+  agx gemini -m gemini-2.0-flash --prompt "hello"
+  agx ollama --model qwen3:8b --prompt "write a poem"
+  agx c --yolo -p "fix the bug"
+
+RAW PASSTHROUGH:
+  Use -- to pass arguments directly to underlying CLI:
+  agx claude -- --resume
+
+NOTE: For predictable LLM usage, always use --prompt or -p flag.`);
+  process.exit(0);
+}
+
+// Resolve provider
+provider = PROVIDER_ALIASES[provider.toLowerCase()];
+if (!provider) {
+  console.error(`Error: Unknown provider "${args[0]}"`);
+  console.error(`Valid providers: ${VALID_PROVIDERS.join(', ')}`);
+  process.exit(1);
+}
+
+const remainingArgs = args.slice(1);
+const translatedArgs = [];
+const rawArgs = [];
+let env = { ...process.env };
+
+// Split raw arguments at --
+const dashIndex = remainingArgs.indexOf('--');
+let processedArgs = remainingArgs;
+if (dashIndex !== -1) {
+  processedArgs = remainingArgs.slice(0, dashIndex);
+  rawArgs.push(...remainingArgs.slice(dashIndex + 1));
+}
+
+// Parsed options (explicit structure for predictability)
+const options = {
+  prompt: null,
+  model: null,
+  yolo: false,
+  print: false,
+  interactive: false,
+  sandbox: false,
+  debug: false,
+  mcp: null
+};
+
+// Collect positional args (legacy support, but --prompt is preferred)
+const positionalArgs = [];
+
+for (let i = 0; i < processedArgs.length; i++) {
+  const arg = processedArgs[i];
+  const nextArg = processedArgs[i + 1];
+
+  switch (arg) {
+    case '--prompt':
+    case '-p':
+      if (nextArg && !nextArg.startsWith('-')) {
+        options.prompt = nextArg;
+        i++;
+      }
+      break;
+    case '--model':
+    case '-m':
+      if (nextArg && !nextArg.startsWith('-')) {
+        options.model = nextArg;
+        i++;
+      }
+      break;
+    case '--yolo':
+    case '-y':
+      options.yolo = true;
+      break;
+    case '--print':
+      options.print = true;
+      break;
+    case '--interactive':
+    case '-i':
+      options.interactive = true;
+      break;
+    case '--sandbox':
+    case '-s':
+      options.sandbox = true;
+      break;
+    case '--debug':
+    case '-d':
+      options.debug = true;
+      break;
+    case '--mcp':
+      if (nextArg && !nextArg.startsWith('-')) {
+        options.mcp = nextArg;
+        i++;
+      }
+      break;
+    default:
+      if (arg.startsWith('-')) {
+        // Unknown flag - pass through
+        translatedArgs.push(arg);
+      } else {
+        // Positional argument (legacy prompt support)
+        positionalArgs.push(arg);
+      }
+  }
+}
+
+// Determine final prompt: explicit --prompt takes precedence
+const finalPrompt = options.prompt || positionalArgs.join(' ') || null;
+
+// Build command based on provider
+let command = '';
+
+// Apply common options to translatedArgs
+if (options.model) {
+  translatedArgs.push('--model', options.model);
+}
+if (options.debug) {
+  translatedArgs.push('--debug');
+}
+
+if (provider === 'gemini') {
+  command = 'gemini';
+
+  // Gemini-specific translations
+  if (options.yolo) translatedArgs.push('--yolo');
+  if (options.sandbox) translatedArgs.push('--sandbox');
+
+  // Gemini prompt handling
+  if (finalPrompt) {
+    if (options.print) {
+      translatedArgs.push('--prompt', finalPrompt);
+    } else if (options.interactive) {
+      translatedArgs.push('--prompt-interactive', finalPrompt);
+    } else {
+      translatedArgs.push(finalPrompt);
+    }
+  }
+} else {
+  // Claude or Ollama
+  command = 'claude';
+
+  // Claude-specific translations
+  if (options.yolo) translatedArgs.push('--dangerously-skip-permissions');
+  if (options.print) translatedArgs.push('--print');
+  if (options.mcp) translatedArgs.push('--mcp-config', options.mcp);
+
+  // Ollama-specific environment setup
+  if (provider === 'ollama') {
+    env.ANTHROPIC_AUTH_TOKEN = 'ollama';
+    env.ANTHROPIC_BASE_URL = 'http://localhost:11434';
+    env.ANTHROPIC_API_KEY = 'none';
+    // Default model for Ollama if not specified
+    if (!options.model) {
+      translatedArgs.push('--model', 'glm-4.7:cloud');
+    }
+  }
+
+  // Claude prompt (positional at end)
+  if (finalPrompt) {
+    translatedArgs.push(finalPrompt);
+  }
+}
+
+// Append raw args at the end
+translatedArgs.push(...rawArgs);
+
+const child = spawn(command, translatedArgs, {
+  env,
+  stdio: 'inherit',
+  shell: false
+});
+
+child.on('exit', (code) => {
+  process.exit(code || 0);
+});
+
+child.on('error', (err) => {
+  if (err.code === 'ENOENT') {
+    console.error(`Error: "${command}" command not found.`);
+  } else {
+    console.error(`Failed to start ${command}:`, err);
+  }
+  process.exit(1);
+});

package/package.json

@@ -0,0 +1,27 @@
+{
+  "name": "@mndrk/agx",
+  "version": "1.0.0",
+  "description": "Unified AI Agent Wrapper for Gemini, Claude, and Ollama",
+  "main": "index.js",
+  "bin": {
+    "agx": "./index.js"
+  },
+  "scripts": {
+    "test": "echo \"Error: no test specified\" && exit 1"
+  },
+  "keywords": [
+    "ai",
+    "agent",
+    "cli",
+    "gemini",
+    "claude",
+    "ollama",
+    "llm"
+  ],
+  "author": "mndrk",
+  "license": "MIT",
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/mndrk/agx"
+  }
+}