mirror of
https://github.com/magnus919/agent-skills.git
synced 2026-09-15 21:46:29 +03:00
Magnus Hedemark
5428b18ddd
Greenfield SkillOpt: 3 epochs for Microsoft AutoGen skill. Conversational agent model, GroupChat patterns, code execution, nested chats, cancellation tokens, MCP integration. All API surfaces validated against microsoft.github.io/autogen docs.
46 lines
1.1 KiB
Markdown
46 lines
1.1 KiB
Markdown
# AutoGen Code Execution
|
|
|
|
## Docker (Recommended)
|
|
|
|
Safe execution of LLM-generated code in isolated containers:
|
|
|
|
```python
|
|
from autogen_ext.code_executors.docker import DockerCommandLineCodeExecutor
|
|
|
|
executor = DockerCommandLineCodeExecutor(work_dir="coding")
|
|
async with executor:
|
|
# Use within GroupChat
|
|
proxy = UserProxyAgent(
|
|
name="proxy",
|
|
code_executor=executor,
|
|
human_input_mode="NEVER",
|
|
)
|
|
```
|
|
|
|
## Local (Development Only)
|
|
|
|
Runs generated code on your machine — use only for trusted environments:
|
|
|
|
```python
|
|
from autogen_ext.code_executors.local import LocalCommandLineCodeExecutor
|
|
|
|
executor = LocalCommandLineCodeExecutor(work_dir="coding")
|
|
```
|
|
|
|
## Cancellation
|
|
|
|
```python
|
|
from autogen_core import CancellationToken
|
|
|
|
token = CancellationToken()
|
|
result = await executor.execute_code_blocks(code_blocks, cancellation_token=token)
|
|
# Cancel via: token.cancel()
|
|
```
|
|
|
|
## Best Practices
|
|
|
|
- Use Docker for any untrusted code execution
|
|
- Set a `work_dir` to isolate generated files
|
|
- Always pass a `CancellationToken` for long-running tasks
|
|
- Monitor `max_turns` to prevent runaway code generation
|