> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tensorstax.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Agent Capabilities
> Learn what TensorStax agents can do and the verification processes they use
## Overview
TensorStax agents are designed to verify every action they take, ensuring reliability and accuracy in your data workflow. The agent provides real-time feedback and validation for all operations it performs.
### dbt Models
**Creation & Editing**: Write new dbt models and edit existing ones with automatic verification.
* Auto-generates preview of model output
* Automatic error detection and fixing
* File versioning for all changes
* Can edit existing dbt models, YML files, tests, and macros
### Tests
**Automated Testing**: Create and execute dbt tests with real-time results.
* Tests auto-execute after creation
* Agent receives immediate pass/fail feedback
* Test results visible to users
* Ensures data quality validation
### YML Files
**Schema Management**: Create and maintain YML configuration files.
* Schema definitions and documentation
* Test configurations
* Model properties and metadata
### Safety Restrictions
**Execution Boundaries**: The agent cannot perform destructive operations.
* **Cannot** execute `dbt build`, `dbt run`, or change tables
* **Cannot** modify database state directly
* Users must press the play button to run models
* Read-only access to database schemas and metadata
### DAG Creation & Editing
**DAG Management**: Create new DAGs and edit existing ones with custom operators.
* Create new Airflow DAGs from scratch
* Edit and modify existing DAG configurations
* Use custom operators tailored to your workflows
* Full DAG lifecycle management
### DAG Simulation
**In-Memory Testing**: Simulate DAG execution without running actual tasks.
* Uses TensorStax custom Airflow compiler
* Simulates DAG execution in memory
* Validates DAG structure and dependencies
* No actual task execution or resource consumption
### Script Creation & Sync
**Glue Script Management**: Create scripts and sync them directly to your AWS environment.
* Create new Glue scripts from scratch
* Automatic sync back to customer's AWS account
* Full integration with AWS Glue service
* Version control and script management
### Interactive Testing
**Live Code Testing**: Test code snippets before finalizing scripts.
* Spin up interactive Glue sessions
* Test code snippets in real-time
* Validate logic before creating final scripts
* Safe testing environment
### AWS Context Integration
**Full AWS Glue Context**: Access to your complete Glue environment.
* Context into Glue connections and databases
* Access to tables, crawlers, and metadata
* Integration with existing Glue resources
* User must press play button to execute scripts
### SELECT Queries & File Management
**Query Execution & Storage**: Run SELECT queries and manage SQL files.
* Execute SELECT queries directly against your database
* Save .sql files in your project for reuse
* Edit existing .sql files with real-time validation
* Preview table changes when editing existing files
### CREATE Statements
**DDL Query Writing**: Write CREATE TABLE and CREATE VIEW statements with preview.
* Write CREATE TABLE and CREATE VIEW queries
* Build on top of existing tables and data
* Preview data output before execution
* Verification against database schema
* User must press play button to execute
### UPDATE & INSERT Statements
**DML Query Writing**: Write UPDATE and INSERT queries with diff preview.
* Write UPDATE queries with change preview
* Incremental inserts from source to type 2 tables
* Data diff preview showing what will change
* No actual table modifications until play button pressed
* Full verification before execution
### Script Execution
**Local Python Processing**: Execute Python scripts on available data.
* Execute regular Python scripts on local CSV files
* Process and analyze data using Python libraries
* Safe execution environment for data exploration
* Integration with existing data sources
### API Structure Analysis
**Pre-Integration Validation**: Check API structures before building data pipelines.
* Analyze API structures and endpoints
* Validate data formats before creating Glue scripts
* Test API responses before building Airflow DAGs
* Ensure compatibility with ingestion workflows
### Environment Variables
**Secure Configuration Management**: Access environment variables stored in Vault.
* Use environment variables set up in HashiCorp Vault
* Secure access to API keys and credentials
* Configuration management for scripts and workflows
* Encrypted variable storage and retrieval
**Verification First**: Every action taken by TensorStax agents includes built-in verification to ensure accuracy and prevent errors from propagating through your data pipeline.