refactor: simplify TaskGuardrail to use LLM for validation, no code generation

This commit is contained in:
Lucas Gomide
2025-04-29 09:50:03 -03:00
parent e3ab80f517
commit e940ff3cbd
14 changed files with 3883 additions and 4217 deletions

View File

@@ -324,11 +324,7 @@ blog_task = Task(
### TaskGuardrail
The `TaskGuardrail` class provides a sophisticated way to generate and execute validation code for task outputs. Here's how it works:
#### Code Execution
The generated code can be executed in two ways: Docker container (Default, Recommended) or current environment (unsafe mode)
The `TaskGuardrail` class offers a robust mechanism for validating task outputs
### Error Handling Best Practices
@@ -791,7 +787,7 @@ task = Task(
)
```
##### Using YAML
#### Using YAML
```yaml
research_task: