Strands Agent Builder

https://github.com/strands-agents/agent-builder
Admin
Strands Agent Builder is an interactive toolkit for building, testing, and extending AI agents and...
Tokens:2,043
Snippets:12
Trust Score:7.8
License:Apache-2.0
Update:1 year ago
Show doc for...
Tokens:
Raw
### Install Strands Agent Builder

Source: https://github.com/strands-agents/agent-builder/blob/main/README.md

Installs the Strands Agent Builder using pipx, a Python application installer.

```bash
pipx install strands-agents-builder
```

--------------------------------

### Run Interactive Agent Development

Source: https://github.com/strands-agents/agent-builder/blob/main/README.md

Starts the interactive command-line interface for developing and managing AI agents.

```bash
strands
```

--------------------------------

### Strands: Custom Model Provider Implementation

Source: https://github.com/strands-agents/agent-builder/blob/main/README.md

Provides an example Python module structure for implementing a custom model provider for Strands Agent Builder. It shows how to expose an 'instance' function that returns a provider object.

```python
$ cat ./.models/custom_model.py
from mymodels import CustomModel

def instance(**config):
    return CustomModel(**config)
```

--------------------------------

### Build and Test Custom Tool

Source: https://github.com/strands-agents/agent-builder/blob/main/README.md

Builds a custom tool named 'sentiment_analyzer' that analyzes text sentiment and tests it with provided examples directly from the command line.

```bash
strands "Create a tool named sentiment_analyzer that analyzes text sentiment and test it with some examples"
```

--------------------------------

### Strands CLI: Knowledge Base Integration

Source: https://github.com/strands-agents/agent-builder/blob/main/README.md

Demonstrates how to load tools and manage agent configurations using the Strands CLI with a specified knowledge base ID. It shows both direct KB ID specification and setting it via an environment variable.

```bash
# Load and extend tools from your knowledge base
strands --kb YOUR_KB_ID "Load my data_visualizer tool and add 3D plotting capabilities"

# Or set a default knowledge base via environment variable
export STRANDS_KNOWLEDGE_BASE_ID="YOUR_KB_ID"
strands "Find my most recent agent configuration and make it more efficient"
```

--------------------------------

### Strands: Model Configuration via Environment Variables

Source: https://github.com/strands-agents/agent-builder/blob/main/README.md

Illustrates how to customize model configuration parameters like maximum tokens and thinking budget using environment variables for the Strands Agent Builder.

```bash
# Maximum tokens for responses
export STRANDS_MAX_TOKENS=32000

# Budget for agent thinking/reasoning
export STRANDS_BUDGET_TOKENS=1024
```

--------------------------------

### Build Agent from Specifications

Source: https://github.com/strands-agents/agent-builder/blob/main/README.md

Builds a specialized agent based on specifications provided in a text file, piped into the strands command.

```bash
cat agent-spec.txt | strands "Build a specialized agent based on these specifications"
```

--------------------------------

### Extend Tool with Knowledge Base

Source: https://github.com/strands-agents/agent-builder/blob/main/README.md

Loads a previously created calculator tool and enhances its functionality with scientific functions by leveraging a knowledge base.

```bash
strands --kb YOUR_KB_ID "Load my previous calculator tool and enhance it with scientific functions"
```

--------------------------------

### Set System Prompt via Environment Variable or File

Source: https://github.com/strands-agents/agent-builder/blob/main/README.md

Demonstrates how to set the system prompt for the Strands Agent Builder. The system prompt can be configured either directly via the STRANDS_SYSTEM_PROMPT environment variable or by writing the prompt content to a `.prompt` file.

```shell
export STRANDS_SYSTEM_PROMPT="You are a Python expert."

# Or local file
echo "You are a security expert." > .prompt
```

--------------------------------

### Strands CLI: Custom Model Provider Configuration

Source: https://github.com/strands-agents/agent-builder/blob/main/README.md

Demonstrates how to configure Strands to use custom model providers via the command line, specifying the provider name and its configuration file or JSON.

```bash
strands --model-provider <NAME> --model-config <JSON|FILE>

# Example using Ollama provider
strands --model-provider ollama --model-config '{"model_id": <ID>}'
```

--------------------------------

### Strands: Default Model Configuration (Bedrock)

Source: https://github.com/strands-agents/agent-builder/blob/main/README.md

Shows the default JSON configuration for the Bedrock model provider, including model ID, token limits, timeouts, and advanced request fields for interleaved and recursive thinking.

```json
{
    "model_id": "us.anthropic.claude-sonnet-4-20250514-v1:0",
    "max_tokens": 32767,
    "boto_client_config": {
        "read_timeout": 900,
        "connect_timeout": 900,
        "retries": {
            "max_attempts": 3,
            "mode": "adaptive"
        }
    },
    "additional_request_fields": {
        "anthropic_beta": ["interleaved-thinking-2025-05-14"],
        "thinking": {
            "type": "enabled",
            "budget_tokens": 2048
        }
    }
}
```

--------------------------------

### Strands Agent Builder Environment Variables

Source: https://github.com/strands-agents/agent-builder/blob/main/README.md

Configuration options for the Strands Agent Builder are managed through environment variables. These variables control model selection, token limits, caching strategies, and other operational aspects of the agent. The table below lists available environment variables, their descriptions, and default values.

```APIDOC
Strands Agent Builder Environment Variables:

STRANDS_MODEL_ID:
  Description: Claude model ID to use for inference.
  Default: us.anthropic.claude-sonnet-4-20250514-v1:0

STRANDS_MAX_TOKENS:
  Description: Maximum tokens for agent responses.
  Default: 32768

STRANDS_BUDGET_TOKENS:
  Description: Token budget for agent thinking/reasoning.
  Default: 2048

STRANDS_THINKING_TYPE:
  Description: Type of thinking capability.
  Default: enabled

STRANDS_ANTHROPIC_BETA:
  Description: Anthropic beta features (comma-separated).
  Default: interleaved-thinking-2025-05-14

STRANDS_CACHE_TOOLS:
  Description: Tool caching strategy.
  Default: default

STRANDS_CACHE_PROMPT:
  Description: Prompt caching strategy.
  Default: default

STRANDS_SYSTEM_PROMPT:
  Description: Custom system prompt (overrides .prompt file).
  Default: None

STRANDS_KNOWLEDGE_BASE_ID:
  Description: Default Knowledge Base ID.
  Default: None

STRANDS_TOOL_CONSOLE_MODE:
  Description: Enable rich console UI.
  Default: enabled

BYPASS_TOOL_CONSENT:
  Description: Skip tool confirmation prompts.
  Default: false
```
Tokens:
Raw