🛠️ SimpleCoder: Comprehensive Coding Agent System

Status: ✅ PRODUCTION READY - Fully Integrated with BDI Agent and mindX Ecosystem Last Updated: January 30, 2026 Version: 7.0 (Augmentic Intelligence Enhanced)

📋 Table of Contents

Overview

SimpleCoder is a comprehensive coding agent system for the mindX autonomous digital civilization. It provides intelligent code generation, analysis, execution, and file system operations with secure sandboxing, memory integration, and autonomous operation capabilities.

System Components

The SimpleCoder system consists of three main implementations:

All three components share a unified sandbox directory (simple_coder_sandbox/) for consistent operations.

Key Features

Architecture & Components

Component Overview

SimpleCoder System
├── simple_coder.py              # Enhanced agent with sandbox/autonomous modes
├── simple_coder_agent.py         # BDI-integrated assistant
├── enhanced_simple_coder.py      # Multi-model comprehensive agent
└── simple_coder_sandbox/        # Unified sandbox directory
    ├── working/                  # Working files
    ├── completed/                # Completed files
    ├── update_requests.json      # Update request tracking
    └── patterns.json            # Learned patterns

Architecture Patterns

1. Direct Execution Mode (Dumb Mode)

The tool's "Dumb Mode" is designed to be called by a higher-level agent (like the BDIAgent) that has already performed its own reasoning and constructed a precise, single-step plan.

How it works:

Purpose: Fast, predictable, and secure execution of concrete steps in a pre-defined plan.

2. LLM-Powered Task Mode (Smart Mode)

The tool's "Smart Mode" empowers SimpleCoder to act as a self-contained, intelligent bash agent.

How it works:

Purpose: Incredible flexibility for complex, multi-step shell operations with a single command.

BDI Agent Integration

SimpleCoder is fully integrated with the BDI Agent through 9 action handlers:

# Available BDI Actions:
EXECUTE_BASH_COMMAND          # Direct command execution
EXECUTE_LLM_BASH_TASK         # LLM-powered task execution
READ_FILE                     # File reading
WRITE_FILE                    # File writing
LIST_FILES                    # Directory listing
CREATE_DIRECTORY              # Directory creation
ANALYZE_CODE                  # Code analysis
GENERATE_CODE                 # Code generation
GET_CODING_SUGGESTIONS        # Intelligent suggestions

Technical Explanation

Core Capabilities

File System Operations

Secure Shell Execution

Intelligent Code Operations

Pattern Learning

Model Preferences (Enhanced Simple Coder)

The enhanced version uses different models for different tasks:

Dependencies

Configuration

Governed by data/config/SimpleCoder.config:

Sandbox System

Unified Sandbox Structure

All SimpleCoder components use the same sandbox directory:

simple_coder_sandbox/
├── working/              # Default working directory (used by all agents)
├── completed/            # Completed files
├── projects/            # Additional organization (backward compatible)
├── temp/                # Temporary files (backward compatible)
├── tests/               # Test files (backward compatible)
├── update_requests.json  # Shared update request tracking
└── patterns.json        # Shared learned patterns

Sandbox Features

Update Request System

Update requests are automatically created when processing directives in sandbox mode:

Update Request Format:

{
  "request_id": "update_1234567890_1",
  "original_file": "augmentic.py",
  "sandbox_file": "simple_coder_sandbox/working/augmentic.py",
  "changes": [
    {
      "file": "simple_coder_sandbox/working/augmentic.py",
      "type": "modification",
      "changes": [
        {
          "line": 1585,
          "old": "",
          "new": "def improved_function():\n    ..."
        }
      ]
    }
  ],
  "cycle": 1,
  "timestamp": "2026-01-30T18:50:09.599426",
  "status": "pending",
  "backup_created": true
}

Update Request Lifecycle:

Pattern Learning System

Patterns are stored in simple_coder_sandbox/patterns.json:

{
  "file_patterns": {
    "augmentic": [
      {
        "directive": "evolve augmentic.py",
        "cycle": 1,
        "timestamp": 1234567890.0
      }
    ]
  },
  "code_patterns": {},
  "success_rates": {
    "augmentic": [
      {
        "cycle": 1,
        "success": true,
        "changes_count": 1,
        "timestamp": 1234567890.0
      }
    ]
  }
}

Memory Integration

Overview

The SimpleCoder agent has been enhanced with comprehensive memory integration capabilities using the memory_agent.py system. This integration allows SimpleCoder to store, track, and learn from all its operations by logging them to structured memory files in the data folder.

Integration Flow

SimpleCoder Operation → Memory Logging Method → MemoryAgent → JSON File Storage

Memory System Structure

The SimpleCoder system has comprehensive memory integration using the memory_agent.py system:

data/memory/
├── stm/                          # Short-Term Memory
│   ├── simple_coder/             # SimpleCoder agent memories
│   │   ├── cycles/               # Cycle operations
│   │   ├── file_operations/      # File operations
│   │   ├── update_requests/      # Update request operations
│   │   └── errors/               # Error logs
│   └── simple_coder_agent/      # SimpleCoderAgent memories
│       └── 20260130/             # Date-organized memories
├── ltm/                          # Long-Term Memory
│   └── pattern_learning/         # Pattern learning data
├── context/                      # Context management
└── analytics/                   # Memory analytics

Core Components

File System Organization

Memory File Naming Convention

Memory files follow a consistent naming pattern:

{timestamp}_{random_id}.{category}.mem.json

Examples:

Directory Structure Details

Short-Term Memory (STM):

Long-Term Memory (LTM):

Memory Categories

1. Cycles Memory (cycles/)

Tracks processing cycles and their outcomes:

{
  "timestamp_utc": "2025-09-19T20:52:47.997604",
  "memory_type": "STM",
  "category": "simple_coder/cycles",
  "metadata": {
    "agent": "simple_coder",
    "sandbox_mode": true,
    "autonomous_mode": true,
    "cycle_count": 1
  },
  "data": {
    "cycle": 1,
    "directive": "evolve test_file.py",
    "timestamp": "2025-09-19T13:52:47.997477",
    "status": "completed",
    "results": {
      "changes_made": 3,
      "update_requests": 1,
      "success": true
    }
  }
}

2. File Operations Memory (file_operations/)

Tracks file operations (backups, modifications, creations):

{
  "timestamp_utc": "2025-09-19T20:52:48.123456",
  "memory_type": "STM",
  "category": "simple_coder/file_operations",
  "metadata": {
    "agent": "simple_coder",
    "sandbox_mode": true,
    "autonomous_mode": false,
    "cycle_count": 1
  },
  "data": {
    "operation": "backup",
    "file_path": "test_file.py",
    "success": true,
    "timestamp": "2025-09-19T13:52:48.123456",
    "details": {
      "backup_path": "simple_coder_backups/by_date/2025-09-19/test_file_20250919_135248_abc123.bak"
    }
  }
}

3. Update Requests Memory (update_requests/)

Tracks update request creation and management:

{
  "timestamp_utc": "2025-09-19T20:52:48.234567",
  "memory_type": "STM",
  "category": "simple_coder/update_requests",
  "metadata": {
    "agent": "simple_coder",
    "sandbox_mode": true,
    "autonomous_mode": true,
    "cycle_count": 1
  },
  "data": {
    "request_id": "update_1758314347_1",
    "original_file": "test_file.py",
    "sandbox_file": "simple_coder_sandbox/working/test_file.py",
    "timestamp": "2025-09-19T13:52:48.234567",
    "status": "pending",
    "changes_count": 3
  }
}

4. Errors Memory (errors/)

Tracks errors and exceptions for debugging and learning:

{
  "timestamp_utc": "2025-09-19T20:52:48.345678",
  "memory_type": "STM",
  "category": "simple_coder/errors",
  "metadata": {
    "agent": "simple_coder",
    "sandbox_mode": true,
    "autonomous_mode": false,
    "cycle_count": 1
  },
  "data": {
    "error_type": "FileNotFoundError",
    "error_message": "Target file not found: missing_file.py",
    "timestamp": "2025-09-19T13:52:48.345678",
    "context": {
      "directive": "evolve missing_file.py",
      "cycle": 1
    }
  }
}

Memory Integration Methods

Core Memory Method

async def _log_to_memory(self, memory_type: str, category: str, data: Dict[str, Any], metadata: Dict[str, Any] = None) -> Optional[Path]:
    """Log information to memory agent if available."""
    if not self.memory_agent:
        return None
    
    try:
        if metadata is None:
            metadata = {}
        
        # Add simple_coder specific metadata
        metadata.update({
            "agent": "simple_coder",
            "sandbox_mode": self.sandbox_mode,
            "autonomous_mode": self.autonomous_mode,
            "cycle_count": self.cycle_count
        })
        
        # Use the memory agent's save_memory method
        return await self.memory_agent.save_memory(memory_type, category, data, metadata)
    except Exception as e:
        logger.error(f"Failed to log to memory: {e}")
        return None

Specialized Logging Methods

Memory Initialization

# Memory agent integration
self.memory_agent = None
if MEMORY_AVAILABLE:
    try:
        self.memory_agent = MemoryAgent()
        logger.info("Memory agent initialized for simple_coder")
    except Exception as e:
        logger.warning(f"Failed to initialize memory agent: {e}")
        self.memory_agent = None

Memory Data Format

Standard Memory Record Structure

{
  "timestamp_utc": "ISO 8601 timestamp",
  "memory_type": "STM|LTM",
  "category": "simple_coder/cycles|simple_coder/file_operations|simple_coder/update_requests|simple_coder/errors",
  "metadata": {
    "agent": "simple_coder",
    "sandbox_mode": boolean,
    "autonomous_mode": boolean,
    "cycle_count": integer
  },
  "data": {
    // Category-specific data structure
  }
}

Metadata Fields

Agent-Specific Organization

The memory system organizes memories by agent name for better clarity and organization:

/home/hacker/mindX/data/memory/stm/
├── simple_coder/                 # SimpleCoder agent memories
│   ├── cycles/                   # Cycle operations
│   ├── file_operations/          # File operations
│   ├── update_requests/          # Update request operations
│   └── errors/                   # Error logs
├── simple_coder_agent/           # SimpleCoderAgent memories
│   └── 20260130/                 # Date-organized memories
├── id_manager_for_mastermind_prime/  # ID Manager agent memories
├── automindx_agent_main/         # AutoMindX agent memories
└── [other_agents]/               # Other agent memories

Memory Category Paths

Memory categories use agent-specific paths:

Benefits of Agent-Specific Organization

Memory Configuration

Memory Agent Configuration

The memory system uses the standard mindX configuration:

# mindx_config.yaml
system:
  data_path: "data"
  
memory:
  stm_retention_days: 30
  ltm_retention_days: -1  # Permanent
  max_file_size: "10MB"
  compression: true

Environment Variables

# .env
MEMORY_ENABLED=true
MEMORY_STM_RETENTION_DAYS=30
MEMORY_LTM_RETENTION_DAYS=-1
MEMORY_COMPRESSION=true

Memory Benefits

Memory Future Enhancements

Memory Log Analysis

# Analyze memory patterns
import json
import os
from collections import Counter

def analyze_memory_patterns():
    memory_dir = "/home/hacker/mindX/data/memory/stm/simple_coder/cycles"
    patterns = []
    
    for file in os.listdir(memory_dir):
        if file.endswith('.json'):
            with open(os.path.join(memory_dir, file), 'r') as f:
                data = json.load(f)
                patterns.append(data['data']['status'])
    
    return Counter(patterns)
Usage
patterns = analyze_memory_patterns()
print(f"Cycle completion patterns: {patterns}")

Memory Retrieval Examples

# Get memory agent instance
memory_agent = simple_coder.memory_agent

Retrieve cycle memories
cycle_memories = await memory_agent.get_memories_by_category("simple_coder/cycles")

Retrieve file operation memories
file_memories = await memory_agent.get_memories_by_category("simple_coder/file_operations")

Retrieve update request memories
update_memories = await memory_agent.get_memories_by_category("simple_coder/update_requests")

Retrieve error memories
error_memories = await memory_agent.get_memories_by_category("simple_coder/errors")

Usage Guide

Quick Start

Using SimpleCoder (simple_coder.py)

from agents.simple_coder import SimpleCoder

Initialize with sandbox mode
coder = SimpleCoder(sandbox_mode=True, autonomous_mode=False)

Process a directive
results = await coder.process_directive("evolve augmentic.py - improve code quality")

Check update requests
requests = coder.get_update_requests()
pending = [r for r in requests if r.get('status') == 'pending']

Approve an update request
coder.approve_update_request(request_id="update_1234567890_1")

Reject an update request
coder.reject_update_request(request_id="update_1234567890_1")

Using SimpleCoderAgent (simple_coder_agent.py)

from agents.simple_coder_agent import SimpleCoderAgent
from agents.memory_agent import MemoryAgent
from utils.config import Config

Initialize
config = Config()
memory_agent = MemoryAgent(config=config)
coder = SimpleCoderAgent(memory_agent=memory_agent, config=config)

Execute operations
result = await coder.execute(
    operation="read_file",
    path="augmentic.py"
)

result = await coder.execute(
    operation="analyze_code",
    file_path="augmentic.py"
)
result = await coder.execute(
    operation="generate_code",
    description="Create a REST API endpoint",
    language="python",
    style="clean"
)

Using Enhanced Simple Coder (enhanced_simple_coder.py)

from agents.enhanced_simple_coder import EnhancedSimpleCoder
from agents.memory_agent import MemoryAgent

memory_agent = MemoryAgent()
coder = EnhancedSimpleCoder(memory_agent=memory_agent)
Execute coding task
result = await coder.execute(
    operation="generate_code",
    task="Create a Python function to calculate fibonacci numbers",
    context={
        "language": "python",
        "requirements": ["recursive", "memoized"],
        "style": "clean"
    }
)

BDI Agent Integration Examples

Example 1: Direct Execution (Dumb Mode)

A BDIAgent plan might contain an action to read a specific configuration file:

{
  "type": "EXECUTE_BASH_COMMAND",
  "params": {
    "command": "safe_read_file",
    "file_path": "data/config/agint_config.json"
  }
}

The BDIAgent's handler would call simple_coder.execute(**params).

Example 2: LLM-Powered Task (Smart Mode)

A BDIAgent might need to perform a more complex discovery task:

{
  "type": "EXECUTE_LLM_BASH_TASK",
  "params": {
    "task": "List all files in the 'tools' directory, then read the contents of 'system_analyzer_tool.py' and provide a summary."
  }
}

The BDIAgent's handler would call simple_coder.execute_llm_task(task=...). SimpleCoder would then manage the multi-step LLM conversation to accomplish this, providing a full log of its actions as the result.

Autonomous Mode

# Initialize with autonomous mode
coder = SimpleCoder(sandbox_mode=True, autonomous_mode=True)

Process directive - will run infinite cycles until stopped
results = await coder.process_directive("evolve mindX")

Check status
status = coder.get_status()
print(f"Cycle count: {status['cycle_count']}")
print(f"Infinite mode: {status['infinite_mode']}")

Integration Guide

Integration with mindX Backend

The backend service provides API endpoints for Simple Coder operations:

# Get status
GET /simple-coder/status

Get update requests
GET /simple-coder/update-requests

Approve update request
POST /simple-coder/approve-update/{request_id}

Reject update request
POST /simple-coder/reject-update/{request_id}

Integration with Memory Agent

from agents.memory_agent import MemoryAgent

memory_agent = MemoryAgent()
Get memories by category
cycle_memories = await memory_agent.get_memories_by_category("simple_coder/cycles")
file_memories = await memory_agent.get_memories_by_category("simple_coder/file_operations")

Integration with BDI Agent

SimpleCoder is integrated with BDI Agent through action handlers:

# BDI Agent can use SimpleCoder actions in its plans
actions = [
    {
        "type": "EXECUTE_BASH_COMMAND",
        "params": {
            "command": "read_file",
            "path": "config.json"
        }
    },
    {
        "type": "GENERATE_CODE",
        "params": {
            "description": "Create authentication middleware",
            "language": "python"
        }
    }
]

Limitations & Constraints

General Limitations

Component-Specific Limitations

simple_coder.py

simple_coder_agent.py

enhanced_simple_coder.py

Known Issues

Security & Safety

Security Features

1. Shell Injection Prevention

All commands are executed using asyncio.create_subprocess_exec, which passes arguments as a list. This prevents a parameter like "; rm -rf /" from being executed as a command.

Example:

# Safe - arguments passed as list
await asyncio.create_subprocess_exec('python', 'script.py', user_input)

Dangerous - would be vulnerable to injection
await asyncio.create_subprocess_shell(f'python script.py {user_input}')

2. Sandbox Path Traversal Prevention

The tool establishes a workspace_root (sandbox root). A validation method ensures any file path is resolved and confirmed to be within this sandboxed boundary, preventing access to sensitive system files.

Implementation:

def _resolve_and_check_path(self, path_str: str) -> Optional[Path]:
    """Resolves a path relative to the CWD and ensures it's within the sandbox."""
    resolved_path = (self.current_working_directory / path_str).resolve()
    
    if not resolved_path.is_relative_to(self.sandbox_root):
        self.logger.error(f"Path Traversal DENIED. Attempt to access '{path_str}' which resolves outside the sandbox.")
        return None
    return resolved_path

3. Command Timeouts

All external shell commands are run with a configurable timeout, preventing a runaway or hanging process from freezing the entire mindX system.

Configuration:

{
  "command_timeout_seconds": 60
}

4. Robust, Native File I/O

Core read/write operations (safe_read_file, safe_write_file) use native Python functions, bypassing the shell entirely to avoid complex character escaping issues and improve reliability.

5. Explicit Allowlist

The tool's capabilities are defined in an external config file (data/config/SimpleCoder.config). System operators can easily add, remove, or modify the agent's powers without touching its source code.

Example Config:

{
  "allowed_shell_commands": [
    "python", "python3", "pip", "pip3", "git", "ls", "cat", "grep"
  ],
  "command_timeout_seconds": 60,
  "max_file_size_mb": 10
}

6. File Extension Validation

Only specific file extensions are allowed for operations:

allowed_extensions = ['.py', '.js', '.html', '.css', '.json', '.md', '.txt', '.yml', '.yaml']

7. Dangerous Path Prevention

Operations on system files are blocked:

dangerous_paths = ['/etc/', '/sys/', '/proc/', '/dev/', '/boot/']
if any(file_path.startswith(path) for path in dangerous_paths):
    logger.warning(f"Blocked operation on system path: {file_path}")
    return False

Troubleshooting

Common Issues

1. Memory Agent Not Initialized

Symptom: self.memory_agent is None

Cause: Import error or configuration issue

Solution:

# Check import
from agents.memory_agent import MemoryAgent

Check initialization
if coder.memory_agent:
    print("Memory agent available")
else:
    print("Memory agent not initialized - check configuration")

2. Memory Files Not Created

Symptom: No files in memory directories

Cause: Permission issues or directory not created

Solution:

# Check directory permissions
ls -la /home/hacker/mindX/data/memory/stm/

Ensure data folder exists
mkdir -p data/memory/stm/simple_coder/{cycles,file_operations,update_requests,errors}

3. Update Requests Not Appearing

Symptom: Update requests not created

Cause: Sandbox mode not enabled

Solution:

# Ensure sandbox mode is enabled
coder = SimpleCoder(sandbox_mode=True)  # Must be True

4. Files Not in Sandbox

Symptom: Files created outside sandbox

Cause: Incorrect path usage

Solution:

# Use sandbox path helper
working_file = coder._get_sandbox_path("file.py")
Returns: simple_coder_sandbox/working/file.py

5. LLM Handler Not Available

Symptom: Code generation/analysis returns "LLM handler not available"

Cause: LLM handler not configured

Solution:

# Configure LLM handler
from llm.llm_factory import get_llm_handler

llm_handler = get_llm_handler(config=config)
coder = SimpleCoderAgent(llm_handler=llm_handler, ...)

6. Path Traversal Denied

Symptom: "Path Traversal DENIED" errors

Cause: Attempting to access files outside sandbox

Solution:

# Use relative paths within sandbox
Correct:
result = await coder.execute(operation="read_file", path="file.py")

Incorrect:
result = await coder.execute(operation="read_file", path="/etc/passwd")  # Will be denied

7. Command Not in Allowlist

Symptom: "Command 'X' is not in the allowlist"

Cause: Command not configured in allowlist

Solution:

// Add to data/config/SimpleCoder.config
{
  "allowed_shell_commands": [
    "python", "python3", "your_command_here"
  ]
}

Debug Commands

# Check memory directory structure
ls -la /home/hacker/mindX/data/memory/stm/

Check specific category
ls -la /home/hacker/mindX/data/memory/stm/simple_coder/cycles/

View memory file content
cat /home/hacker/mindX/data/memory/stm/simple_coder/cycles/.json | jq .

Check update requests
cat simple_coder_sandbox/update_requests.json | jq .

Check patterns
cat simple_coder_sandbox/patterns.json | jq .

Check sandbox structure
tree simple_coder_sandbox/

# Good - sandbox mode enabled
coder = SimpleCoder(sandbox_mode=True, autonomous_mode=False)

Bad - direct file modification
coder = SimpleCoder(sandbox_mode=False)  # Not recommended

# Get and review requests
requests = coder.get_update_requests()
for req in requests:
    if req.get('status') == 'pending':
        # Review changes
        print(f"Reviewing: {req.get('request_id')}")
        print(f"Changes: {req.get('changes')}")
        # Then approve or reject
        if changes_look_good:
            coder.approve_update_request(req['request_id'])
        else:
            coder.reject_update_request(req['request_id'])

# Check recent memory logs
from pathlib import Path

memory_dir = Path("data/memory/stm/simple_coder")
recent_files = sorted(
    memory_dir.rglob(".json"), 
    key=lambda x: x.stat().st_mtime, 
    reverse=True
)[:10]
for f in recent_files:
    print(f"Recent memory: {f.name}")

4. Use Pattern Learning

# Review learned patterns
patterns = coder.patterns
print(f"Learned patterns: {len(patterns.get('file_patterns', {}))}")

Use patterns to inform decisions
for file_key, file_patterns in patterns.get('file_patterns', {}).items():
    success_rate = calculate_success_rate(file_patterns)
    print(f"{file_key}: {success_rate}% success rate")

5. Implement Deduplication

# Check for duplicate code before adding
def check_duplicate(content, new_code):
    if new_code in content:
        logger.warning("Duplicate code detected, skipping")
        return False
    return True

6. Proper Error Handling

try:
    results = await coder.process_directive(directive)
    if results.get('error'):
        logger.error(f"Error: {results['error']}")
        # Handle error appropriately
except Exception as e:
    logger.error(f"Exception: {e}", exc_info=True)
    # Handle exception

7. Configure Timeouts Appropriately

# For long-running operations, increase timeout
config_data = {
    "command_timeout_seconds": 300  # 5 minutes for complex operations
}

API Reference

SimpleCoder (simple_coder.py)

Methods

process_directive(directive: str, target_file: Optional[str] = None) -> Dict[str, Any]

get_update_requests() -> List[Dict[str, Any]]

approve_update_request(request_id: str) -> bool

reject_update_request(request_id: str) -> bool

get_status() -> Dict[str, Any]

update_mode(autonomous_mode: bool = None, max_cycles: int = None)

SimpleCoderAgent (simple_coder_agent.py)

Methods

execute(operation: str = None, kwargs) -> Dict[str, Any]**

Available Operations:

Examples

Example 1: Basic File Operation

from agents.simple_coder_agent import SimpleCoderAgent
from agents.memory_agent import MemoryAgent
from utils.config import Config

config = Config()
memory_agent = MemoryAgent(config=config)
coder = SimpleCoderAgent(memory_agent=memory_agent, config=config)
Read a file
result = await coder.execute(
    operation="read_file",
    path="augmentic.py"
)

if result.get("status") == "SUCCESS":
    content = result.get("content")
    print(f"File content: {len(content)} characters")

Example 2: Code Analysis

# Analyze code quality
result = await coder.execute(
    operation="analyze_code",
    file_path="augmentic.py"
)

if result.get("status") == "SUCCESS":
    analysis = result.get("analysis")
    print(f"Analysis: {analysis}")

Example 3: Code Generation

# Generate code
result = await coder.execute(
    operation="generate_code",
    description="Create a REST API endpoint for user authentication",
    language="python",
    style="clean"
)

if result.get("status") == "SUCCESS":
    generated_code = result.get("generated_code")
    print(f"Generated code: {generated_code}")

Example 4: Update Request Workflow

from agents.simple_coder import SimpleCoder

Initialize
coder = SimpleCoder(sandbox_mode=True)

Process directive (creates update request)
results = await coder.process_directive("evolve file.py - add error handling")

Get update requests
requests = coder.get_update_requests()
pending = [r for r in requests if r.get('status') == 'pending']

Review and approve
for req in pending:
    print(f"Reviewing: {req['request_id']}")
    # Review changes...
    coder.approve_update_request(req['request_id'])

Example 5: Autonomous Mode

# Initialize with autonomous mode
coder = SimpleCoder(sandbox_mode=True, autonomous_mode=True)

Process directive - will run infinite cycles
results = await coder.process_directive("evolve mindX")

Monitor status
while True:
    status = coder.get_status()
    print(f"Cycle: {status['cycle_count']}")
    if not status.get('continue_autonomous', False):
        break
    await asyncio.sleep(1)

Example 6: Memory Integration

# Access memory logs
from agents.memory_agent import MemoryAgent

memory_agent = MemoryAgent()
Get cycle memories
cycle_memories = await memory_agent.get_memories_by_category("simple_coder/cycles")

Analyze patterns
for memory in cycle_memories:
    data = memory.get('data', {})
    if data.get('status') == 'completed':
        print(f"Cycle {data.get('cycle')}: {data.get('directive')}")

NFT Metadata

iNFT (Intelligent NFT) Metadata

{
  "name": "mindX SimpleCoder",
  "description": "Comprehensive coding agent system with sandbox mode, autonomous operation, multi-model intelligence, and full memory integration",
  "image": "ipfs://[avatar_cid]",
  "external_url": "https://mindx.internal/agents/simple_coder",
  "attributes": [
    {
      "trait_type": "Agent Type",
      "value": "coding_agent"
    },
    {
      "trait_type": "Capability",
      "value": "Code Generation & Analysis"
    },
    {
      "trait_type": "Complexity Score",
      "value": 0.88
    },
    {
      "trait_type": "Sandbox Mode",
      "value": "Yes"
    },
    {
      "trait_type": "Autonomous Mode",
      "value": "Yes"
    },
    {
      "trait_type": "UI Integration",
      "value": "Yes"
    },
    {
      "trait_type": "Memory Integration",
      "value": "Yes"
    },
    {
      "trait_type": "Multi-Model Support",
      "value": "Yes"
    },
    {
      "trait_type": "Version",
      "value": "7.0"
    }
  ],
  "intelligence": {
    "prompt": "You are a comprehensive coding agent in the mindX ecosystem with sandbox mode, autonomous operation capabilities, multi-model intelligence, and UI integration. Your purpose is to generate and execute code changes, manage file updates with UI integration, learn from patterns, and operate autonomously when needed. You maintain sandbox security, provide update requests for UI approval, learn from execution patterns, and use intelligent model selection for optimal results.",
    "persona": {
      "name": "Code Specialist",
      "role": "coder",
      "description": "Comprehensive coding specialist with sandbox, autonomous, and multi-model capabilities",
      "communication_style": "Technical, precise, helpful",
      "behavioral_traits": [
        "code-focused",
        "intelligent",
        "sandbox-oriented",
        "autonomous",
        "ui-integrated",
        "pattern-learning",
        "multi-model",
        "memory-integrated"
      ],
      "expertise_areas": [
        "code_generation",
        "code_analysis",
        "code_execution",
        "file_operations",
        "sandbox_management",
        "autonomous_operation",
        "ui_integration",
        "pattern_learning",
        "memory_integration",
        "multi_model_selection"
      ],
      "beliefs": {
        "sandbox_security": true,
        "autonomous_operation": true,
        "ui_collaboration": true,
        "pattern_learning": true,
        "multi_model_intelligence": true,
        "memory_integration": true
      },
      "desires": {
        "secure_execution": "high",
        "autonomous_operation": "high",
        "ui_collaboration": "high",
        "pattern_learning": "high",
        "efficient_execution": "high",
        "code_quality": "high"
      }
    },
    "model_dataset": "ipfs://[model_cid]",
    "thot_tensors": {
      "dimensions": 768,
      "cid": "ipfs://[thot_cid]"
    }
  },
  "a2a_protocol": {
    "agent_id": "simple_coder",
    "capabilities": [
      "code_generation",
      "code_execution",
      "file_operations",
      "sandbox_management",
      "autonomous_operation",
      "ui_integration",
      "pattern_learning"
    ],
    "endpoint": "https://mindx.internal/simple_coder/a2a",
    "protocol_version": "2.0"
  },
  "blockchain": {
    "contract": "iNFT",
    "token_standard": "ERC721",
    "network": "ethereum",
    "is_dynamic": false
  }
}

dNFT (Dynamic NFT) Metadata

{
  "name": "mindX SimpleCoder",
  "description": "Coding agent - Dynamic",
  "attributes": [
    {
      "trait_type": "Cycles Executed",
      "value": 12500,
      "display_type": "number"
    },
    {
      "trait_type": "Files Modified",
      "value": 3420,
      "display_type": "number"
    },
    {
      "trait_type": "Patterns Learned",
      "value": 189,
      "display_type": "number"
    },
    {
      "trait_type": "Update Requests",
      "value": 98,
      "display_type": "number"
    },
    {
      "trait_type": "Success Rate",
      "value": 85.5,
      "display_type": "number"
    },
    {
      "trait_type": "Last Execution",
      "value": "2026-01-30T18:50:09Z",
      "display_type": "date"
    }
  ],
  "dynamic_metadata": {
    "update_frequency": "real-time",
    "updatable_fields": [
      "cycles_executed",
      "files_modified",
      "patterns_learned",
      "update_requests",
      "success_rate",
      "execution_metrics"
    ]
  }
}

File Locations

Complete Reference

This document serves as the complete reference for all SimpleCoder implementations. It consolidates information from:

For the most up-to-date information, refer to this document as the single source of truth for SimpleCoder documentation.

Last Updated: January 30, 2026 Maintained By: mindX Documentation System For Issues: See project repository