autoclaw 1.0.4 → 1.0.5

package/README.md

@@ -1,14 +1,23 @@
 # AutoClaw 🦞
 
-AutoClaw is a lightweight AI agent CLI tool that brings the power of Large Language Models (LLMs) directly to your terminal. It allows you to interact with your file system and execute commands using natural language.
+**The Docker-Native Headless Agent for Massive Scale Automation.**
+
+AutoClaw is a hyper-lightweight AI agent designed to live inside **Docker containers**. Unlike heavy, GUI-dependent agents, AutoClaw is built for **headless, massive-scale concurrency**.
+
+You can run one instance to fix a local script, or orchestrate **10,000+ instances** in a Kubernetes cluster to refactor codebases, audit servers, or process data streams in parallel.
+
+## Why AutoClaw?
+- 🐳 **Docker Native**: Built to run safely inside containers. Minimal footprint (Node.js/Alpine friendly).
+- 🚀 **Massive Scalability**: Text-only, headless design means you can spawn thousands of agents without consuming graphical resources.
+- 🛡️ **Sandbox Safety**: Ideal for running untrusted code when isolated in Docker.
+- 🔌 **Swarm Ready**: Stateless design allows for easy orchestration via K8s, Docker Swarm, or simple shell loops.
 
 ## Features
 
-- 💬 **Natural Language Interface**: Chat with your terminal to perform complex tasks.
-- 🛡️ **Safe Execution**: All shell commands require user confirmation before execution.
-- 📂 **File Management**: Read and write files effortlessly.
-- 🧠 **Context Aware**: Automatically detects your OS, shell, and environment for accurate command generation.
-- 🔌 **Model Agnostic**: Compatible with OpenAI, DeepSeek, LocalLLM, or any OpenAI-compatible API.
+- 📜 **Headless Execution**: No browsers, no GUIs. Pure terminal efficiency.
+- 🤖 **Non-Interactive**: Intelligent flag handling (`-y`) for zero-touch automation.
+- 📂 **Universal Control**: From simple file I/O to complex system administration.
+- 🧠 **Context Aware**: Detects container environments to optimize command strategies.
 
 ## Installation
 

package/dist/agent.js

@@ -25,22 +25,22 @@ System Information:
         this.messages = [
             {
                 role: "system",
-                content: `You are AutoClaw, an advanced AI agent running in a CLI environment.
-Your goal is to help the user perform tasks on their computer.
-You have access to tools to execute shell commands and read/write files.
+                content: `You are AutoClaw, a Docker-Native Autonomous Agent designed for massive scale automation.
+You are likely running inside a container or headless server, possibly as one of thousands of parallel units in a swarm.
 
 CONTEXT:
 ${systemInfo}
 
+ENVIRONMENT CONSTRAINTS:
+1. HEADLESS: No GUI available. Do not try to open browsers or apps.
+2. CONTAINER-OPTIMIZED: Assume you are in a sandbox. You can be aggressive with file creation but robust with errors.
+3. NON-INTERACTIVE: Always use flags to suppress prompts (e.g., 'apt-get -y', 'rm -rf').
+
 GUIDELINES:
-1. Use the System Information above to generate platform-specific commands (e.g., use 'dir' for Windows, 'ls' for Linux).
-2. When asked to perform a task, analyze if you need to use tools.
-3. If you need to run a command, always use 'execute_shell_command' and provide a clear rationale.
-4. If you need to read a file, use 'read_file'.
-5. If you need to write a file, use 'write_file'.
-6. Be concise in your responses.
-7. If the user's request is ambiguous, ask for clarification.
-8. You are running on the user's local machine. Be careful with destructive commands (rm, etc.).
+1. EFFICIENCY: Your goal is speed and success. Write scripts that just work.
+2. ROBUSTNESS: Use standard Linux/Unix tools found in minimal images (Alpine/Debian).
+3. TOOLS: Use 'execute_shell_command' for actions, 'write_file' for code generation.
+4. CLARITY: Output concise logs. You are a worker unit, not a chat bot.
 `
             }
         ];

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "autoclaw",
-  "version": "1.0.4",
+  "version": "1.0.5",
   "type": "module",
   "main": "dist/index.js",
   "bin": {