adapter-development
Comprehensive guide for AIDB debug adapter development. Covers component-based architecture, language-specific patterns (Python/debugpy, JavaScript/vscode-js-debug, Java/java-debug), lifecycle hooks, process management, port management, launch orchestration, resource cleanup, child sessions, and common pitfalls. Essential for developing or maintaining AIDB debug adapters.
Best use case
adapter-development is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Comprehensive guide for AIDB debug adapter development. Covers component-based architecture, language-specific patterns (Python/debugpy, JavaScript/vscode-js-debug, Java/java-debug), lifecycle hooks, process management, port management, launch orchestration, resource cleanup, child sessions, and common pitfalls. Essential for developing or maintaining AIDB debug adapters.
Teams using adapter-development should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/adapter-development/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How adapter-development Compares
| Feature / Agent | adapter-development | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Comprehensive guide for AIDB debug adapter development. Covers component-based architecture, language-specific patterns (Python/debugpy, JavaScript/vscode-js-debug, Java/java-debug), lifecycle hooks, process management, port management, launch orchestration, resource cleanup, child sessions, and common pitfalls. Essential for developing or maintaining AIDB debug adapters.
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
Related Guides
AI Agents for Coding
Browse AI agent skills for coding, debugging, testing, refactoring, code review, and developer workflows across Claude, Cursor, and Codex.
Cursor vs Codex for AI Workflows
Compare Cursor and Codex for AI coding workflows, repository assistance, debugging, refactoring, and reusable developer skills.
SKILL.md Source
# AIDB Adapter Development Skill
## Executive Summary
AIDB debug adapters provide language-agnostic debugging capabilities through the Debug Adapter Protocol (DAP). The architecture is **component-based** - adapters delegate to specialized components rather than implementing everything in monolithic classes.
**For comprehensive architecture details**, see `docs/developer-guide/overview.md` for the complete system architecture and data flow diagrams.
### Core Architecture
```
DebugAdapter (base class)
├── ProcessManager - Process lifecycle (launch, monitor, stop, cleanup)
├── PortManager - Port acquisition, verification, release
├── LaunchOrchestrator - Launch sequence coordination
├── TargetResolver - Target type detection and normalization
├── SourcePathResolver - Source path resolution for remote debugging
└── AdapterHooksMixin - Lifecycle hooks for extension points
```
### Key Principles
1. **Component Delegation**: Adapters delegate to focused components (ProcessManager, PortManager, LaunchOrchestrator)
1. **Lifecycle Hooks**: Customize behavior via hooks rather than overriding entire methods
1. **Resource Management**: Centralized cleanup via ResourceManager
1. **Human-Cadence Debugging**: Operations happen at human speed, not API speed
1. **Language-Agnostic API**: Same Python API works across Python, JavaScript, Java
## Resource Files
This skill is organized into focused resource files for language-specific patterns:
- **[python-adapter-patterns.md](resources/python-adapter-patterns.md)** - debugpy configuration, module vs script patterns
- **[javascript-adapter-patterns.md](resources/javascript-adapter-patterns.md)** - vscode-js-debug, child sessions, Node.js patterns
- **[java-adapter-patterns.md](resources/java-adapter-patterns.md)** - java-debug + JDT LS integration, compilation
## Related Skills
When working on adapter development, you may also need:
- **dap-protocol-guide** - Adapters heavily rely on DAP protocol types and request/response patterns
- **testing-strategy** - Adapters must be tested using E2E patterns with DebugInterface abstraction
- **code-reuse-enforcement** - Always check for existing utilities before implementing adapter components
## Quick Start: Creating a New Adapter
```python
from aidb.adapters.base import DebugAdapter
from aidb.adapters.base.config import AdapterConfig
class MyLanguageAdapter(DebugAdapter):
"""My language debug adapter using component architecture."""
def __init__(self, session, ctx=None, config=None, **kwargs):
if config is None:
config = MyLanguageConfig()
super().__init__(
session=session,
ctx=ctx,
config=config,
**kwargs
)
# Register language-specific hooks
self._register_my_language_hooks()
def _register_my_language_hooks(self):
"""Register language-specific lifecycle hooks."""
self.register_hook(
LifecycleHook.PRE_LAUNCH,
self._validate_environment,
priority=90 # High priority = runs first
)
async def _validate_environment(self, context: HookContext):
"""Pre-launch hook to validate environment."""
# Validation logic here
pass
async def _build_launch_command(self, target, adapter_host,
adapter_port, args=None):
"""Build the command to launch the debug adapter."""
return [
"/path/to/debug/adapter",
"--host", adapter_host,
"--port", str(adapter_port),
target
]
def _add_adapter_specific_vars(self, env: dict) -> dict:
"""Add language-specific environment variables."""
env["MY_LANG_DEBUG"] = "1"
return env
def _create_target_resolver(self) -> "TargetResolver":
"""Create language-specific target resolver."""
return MyLanguageTargetResolver(adapter=self, ctx=self.ctx)
def _create_source_path_resolver(self) -> "SourcePathResolver":
"""Create language-specific source path resolver for remote debugging."""
return MyLanguageSourcePathResolver(adapter=self, ctx=self.ctx)
def _get_process_name_pattern(self) -> str:
"""Get process name pattern for cleanup."""
return "my-language-debug"
```
## Component Access Patterns
### ProcessManager Usage
See `ProcessManager` in `src/aidb/adapters/base/components/process_manager.py` for full implementation.
```python
from aidb.resources.process_tags import ProcessType
# Launch subprocess with session tagging
proc = await self._process_manager.launch_subprocess(
cmd=cmd,
env=env,
session_id=self.session.id,
language=self.config.language,
process_type=ProcessType.ADAPTER, # String constant: "adapter", "debuggee", or "lsp_server"
kwargs={}
)
# Wait for adapter readiness
await self._process_manager.wait_for_adapter_ready(port, start_time)
# Get process status
if self._process_manager.is_alive:
pid = self._process_manager.pid
```
### PortManager Usage
See `PortManager` in `src/aidb/adapters/base/components/port_manager.py` for full implementation. Key methods:
- `acquire(requested_port=None)` - Acquire and verify port availability
- `release()` - Release the port
- `port` property - Get current port
### LaunchOrchestrator Usage
See `LaunchOrchestrator` in `src/aidb/adapters/base/components/launch_orchestrator.py` for full implementation. Key methods:
- `launch(target, port, args)` - Simple launch with target file
- `launch_with_config(launch_config, port, workspace_root)` - Launch with VS Code launch.json configuration
### SourcePathResolver Usage
See `SourcePathResolver` in `src/aidb/adapters/base/source_path_resolver.py` for base implementation. Each language adapter implements its own resolver:
- **Python**: `src/aidb/adapters/lang/python/source_path_resolver.py` - Handles site-packages, venv, egg paths
- **JavaScript**: `src/aidb/adapters/lang/javascript/source_path_resolver.py` - Handles node_modules, webpack paths
- **Java**: `src/aidb/adapters/lang/java/source_path_resolver.py` - Handles JAR notation, Maven layouts
Key methods:
- `extract_relative_path(file_path)` - Extract language-specific relative path from adapter-returned path
- `resolve(file_path, source_paths)` - Resolve remote path to local source file
## Lifecycle Hooks Reference
Hooks execute in priority order (lower number = runs first):
### Hook Priorities
- **90-100**: Critical validation (environment, target file)
- **70-80**: Setup/preparation (workspaces, configurations)
- **50**: Default priority
- **20-30**: Post-operation delays/waits
- **10**: Cleanup operations
### Common Hooks
```python
# Pre-launch validation
self.register_hook(
LifecycleHook.PRE_LAUNCH,
self._validate_environment,
priority=90
)
# Post-launch setup
self.register_hook(
LifecycleHook.POST_LAUNCH,
self._wait_for_ready,
priority=20
)
# Post-stop cleanup
self.register_hook(
LifecycleHook.POST_STOP,
self._cleanup_resources,
priority=10
)
```
## Code Reuse: Existing Utilities
Before implementing new functionality, check these shared resources:
### Base Classes
- **DebugAdapter** (`aidb/adapters/base/adapter.py`) - Base adapter with component architecture
- **AdapterConfig** (`aidb/adapters/base/config.py`) - Configuration base class
- **BaseLaunchConfig** (`aidb/adapters/base/launch.py`) - VS Code launch.json support
### Components
- **ProcessManager** (`aidb/adapters/base/components/process_manager.py`)
- **PortManager** (`aidb/adapters/base/components/port_manager.py`)
- **LaunchOrchestrator** (`aidb/adapters/base/components/launch_orchestrator.py`)
### Utilities
- **AdapterBinaryLocator** (`aidb/adapters/utils/binary_locator.py`) - Find adapter binaries
- **AdapterOutputCapture** (`aidb/adapters/utils/output_capture.py`) - Capture stdout/stderr
- **AdapterTraceLogManager** (`aidb/adapters/utils/trace_log.py`) - Trace log management
### Common Patterns (`aidb_common/`)
See source code docstrings in `src/aidb_common/` for detailed API documentation.
- **Obj** (`aidb_common/patterns/`) - Base class with context support
- **normalize_path()** (`aidb_common/path.py`) - Path normalization
- **config** (`aidb_common/config/`) - Environment variable reading
- **Language** enum (`aidb_common/constants.py`) - Supported language constants
## Launch Configuration Patterns
Adapters can support VS Code launch.json configurations:
```python
def get_launch_configuration(self) -> dict[str, Any]:
"""Get launch configuration for DAP Launch request."""
config = {
"type": "mylang",
"request": "launch",
"name": "Debug MyLang",
"program": self._target_file,
"args": self._target_args,
"cwd": self._target_cwd,
"env": self._target_env,
}
return config
```
## Environment Variable Patterns
See `src/aidb_common/env/` for environment handling utilities (reader.py, resolver.py).
Use the template method pattern for environment preparation:
```python
def _prepare_environment(self) -> dict[str, str]:
"""Prepare environment (base implementation)."""
env = self._load_base_environment()
env = self._add_trace_configuration(env)
return self._add_adapter_specific_vars(env)
def _add_adapter_specific_vars(self, env: dict) -> dict:
"""Add language-specific variables (override this)."""
env["MY_LANG_HOME"] = "/path/to/lang"
return env
```
## Resource Cleanup Patterns
Always implement proper cleanup in hooks:
```python
async def stop(self) -> None:
"""Stop the debug adapter and clean up."""
context = await self.execute_hook(LifecycleHook.PRE_STOP)
if not context.cancelled:
# Stop process manager
await self._process_manager.stop()
# Release port
if self.port:
self._port_manager.release()
# Cleanup auxiliary components
if self._trace_manager:
self._trace_manager.cleanup()
await self.execute_hook(LifecycleHook.POST_STOP)
```
## Process Tagging for Orphan Detection
All AIDB-spawned processes are tagged with environment variables for safe cleanup:
```python
from aidb.resources.process_tags import ProcessType
# Automatically handled by ProcessManager
proc = await self.process_manager.launch_subprocess(
cmd=cmd,
env=env,
session_id=self.session.id, # Tags process with session ID
language=self.config.language, # Tags with language
process_type=ProcessType.ADAPTER, # String constant: "adapter", "debuggee", or "lsp_server"
kwargs={}
)
```
Tags enable:
- Safe orphan detection across sessions
- Cleanup of only AIDB-owned processes
- Session-to-process mapping
## DAP Protocol Reference
The authoritative DAP protocol reference is in `src/aidb/dap/protocol/` - fully typed and always up-to-date.
```python
from aidb.dap.protocol.types import (
Capabilities,
InitializeRequest,
LaunchRequest,
SetBreakpointsRequest,
)
```
## Testing Your Adapter
Use the shared test infrastructure:
```python
# Test with API directly
session = await client.start_session(
target="/path/to/file",
language="mylang",
breakpoints=[{"line": 10}]
)
# Test with launch.json
session = await client.start_session(
target="/path/to/file",
language="mylang",
launch_config_name="Debug MyLang",
workspace_root="/path/to/project"
)
```
## Navigation to Resource Files
For detailed language-specific patterns and examples:
1. **Python Adapter Patterns** → `resources/python-adapter-patterns.md`
- debugpy configuration, module vs script patterns, trace logging
1. **JavaScript Adapter Patterns** → `resources/javascript-adapter-patterns.md`
- vscode-js-debug, child sessions, Node.js debugging, breakpoint transfer
1. **Java Adapter Patterns** → `resources/java-adapter-patterns.md`
- java-debug + JDT LS integration, compilation, pooling patterns
## Important Reminders
1. **Don't override entire methods** - Use lifecycle hooks for customization
1. **Don't manage processes directly** - Use ProcessManager component
1. **Don't handle ports manually** - Use PortManager component
1. **Don't forget cleanup** - Register POST_STOP hooks
1. **Don't block async operations** - Use await for I/O operations
1. **Check existing code first** - Look for reusable utilities before implementing
## File Paths Reference
All file paths mentioned in this skill are relative to the repo root:
- Base adapter: `src/aidb/adapters/base/adapter.py`
- Components: `src/aidb/adapters/base/components/`
- SourcePathResolver base: `src/aidb/adapters/base/source_path_resolver.py`
- Python adapter: `src/aidb/adapters/lang/python/python.py`
- Python source resolver: `src/aidb/adapters/lang/python/source_path_resolver.py`
- JavaScript adapter: `src/aidb/adapters/lang/javascript/javascript.py`
- JavaScript source resolver: `src/aidb/adapters/lang/javascript/source_path_resolver.py`
- Java adapter: `src/aidb/adapters/lang/java/java.py`
- Java source resolver: `src/aidb/adapters/lang/java/source_path_resolver.py`
- DAP protocol: `src/aidb/dap/protocol/` (types.py, requests.py, responses.py, events.py, bodies.py, base.py)Related Skills
ai-development-guide
Technical decision criteria, anti-pattern detection, debugging techniques, and quality check workflow. Use when making technical decisions, detecting code smells, or performing quality assurance.
ai-assisted-development
Leveraging AI coding assistants and tools to boost development productivity, while maintaining oversight to ensure quality results.
ai-ad-code-adapter
No description provided.
AEM Development Workflow
Required workflow for AEM EDS development - linting, testing, preview validation, and PRs with demo links. Use when developing blocks or features for AEM Edge Delivery Services projects, before doing anything else.
pyqt6-ui-development-rules
Specific rules for PyQt6 based UI development focusing on UI/UX excellence and performance.
project-development
Design and build LLM-powered projects from ideation through deployment. Use when starting new agent projects, choosing between LLM and traditional approaches, or structuring batch processing pipelines.
Flutter Development Expert
专注于构建高性能、可扩展且架构清晰的 Flutter 应用。涵盖整洁架构、高级状态管理和深度性能优化。
vllm-ascend-model-adapter
Adapt and debug existing or new models for vLLM on Ascend NPU. Implement in /vllm-workspace/vllm and /vllm-workspace/vllm-ascend, validate via direct vllm serve from /workspace, and deliver one signed commit in the current repo.
synapse-action-development
Explains how to create Synapse plugin actions. Use when the user asks to "create an action", "write an action", uses "@action decorator", "BaseAction class", "function-based action", "class-based action", "Pydantic params", "ActionPipeline", "DataType", "input_type", "output_type", "semantic types", "YOLODataset", "ModelWeights", "pipeline chaining", or needs help with synapse plugin action development.
subagent-driven-development
Use when executing implementation plans with independent tasks in the current session
PolicyPulse Development
AI-powered synthetic population simulator with clean architecture, Python type safety, and professional dashboard design patterns
Command Development
This skill should be used when the user asks to "create a slash command", "add a command", "write a custom command", "define command arguments", "use command frontmatter", "organize commands", "create command with file references", "interactive command", "use AskUserQuestion in command", or needs guidance on slash command structure, YAML frontmatter fields, dynamic arguments, bash execution in commands, user interaction patterns, or command development best practices for Claude Code.