V0.0.4 (#11)

* better parser for the output
* refine prompt for the plan summary

.header.md

@@ -2,42 +2,64 @@
 
 ## Overview
 
-Integrate Amazon Bedrock to your HashiCorp Cloud Platform Terraform (Terraform Cloud) Run Tasks for:
+Enhance your HashiCorp Cloud Platform Terraform (Terraform Cloud) workflows with AI-powered analysis using Amazon Bedrock. This module integrates seamlessly as a Run Task to provide:
 
-* Analyzing Terraform plan and generate short-summary
+* Intelligent Terraform plan analysis with concise, human-friendly summaries
+* Advanced function calling capabilities for extended analysis (e.g., AMI evaluation)
+* Responsible AI implementation with customizable guardrails
 
-* Function calling for other API-based analysis (e.g AMI analysis)
+### Key Features
 
-### Terraform plan summary
+1. **AI-Powered Plan Summaries**
+   * Generate clear, natural language summaries of Terraform plan outputs
+   * Quickly understand the impact of infrastructure changes
+   ![Example](./images/example.png)
 
-Summarize Terraform plan output in human friendly natural language.
-![Example](./images/example.png)
+2. **Extensible Analysis with Function Calling**
+   * Leverage AI to perform additional analyses, such as AMI evaluation
+   * Easily extend to other API-based tools and services
+   ![Example2](./images/example2.png)
 
-### Function calling (AMI analysis)
+3. **Responsible AI Integration**
+   * Implement Amazon Bedrock guardrails tailored to your organization's policies
+   * Ensure ethical and compliant AI usage in your infrastructure workflows
+   ![Example3](./images/example3.png)
 
-Use function calling to execute other tools, such as analyzing AMI data.
-![Example2](./images/example2.png)
+4. **Secure Architecture**
+   * Designed for deployment in a dedicated AWS account
+   * Optional AWS WAF integration for enhanced endpoint protection
+   * Adherence to AWS security best practices
 
-### Responsible AI
+5. **Seamless Terraform Cloud Integration**
+   * Operates as a native Run Task within your Terraform Cloud workflow
+   * Provides insights without disrupting existing processes
 
-Implement safeguards using Amazon Bedrock guardrails customized to your organization requirements and responsible AI policies
-![Example3](./images/example3.png)
+6. **Flexible and Customizable**
+   * Adapt the module to fit your specific organizational needs
+   * Easily configure and extend guardrails and analysis capabilities
 
 ## Architecture
 
 ![Diagram](./images/arch.png)
 
-This module must be deployed in AWS account with access to Amazon Bedrock model. We recommend you to use hub-spoke model, by deploying this module in one dedicated account.
-
-Please refer to the [best-practice](#best-practice) section below for more details.
+This module leverages a hub-spoke model, designed for deployment in a dedicated AWS account with Amazon Bedrock access. It utilizes AWS Lambda, CloudFront, and other AWS services to provide a scalable and secure solution.
 
 ## Prerequisites
 
-To use this module you need have the following:
+To implement this module, you'll need:
+
+1. An AWS account with appropriate credentials
+2. Access to Amazon Bedrock (default model: Claude 3 Sonnet)
+3. A HashiCorp Cloud Platform (HCP) Terraform account
+
+## Getting Started
+
+For detailed setup instructions and best practices, please refer to the sections below:
+
+* [Usage](#usage)
+* [Best Practices](#best-practice)
 
-1. AWS account and credentials
-1. Amazon Bedrock model access (default model is `Claude 3 Sonnet`)
-1. HCP Terraform account
+Enhance your Terraform workflows with AI-powered insights while maintaining security and responsible AI practices.
 
 ## Usage
 

README.md

@@ -3,42 +3,64 @@
 
 ## Overview
 
-Integrate Amazon Bedrock to your HashiCorp Cloud Platform Terraform (Terraform Cloud) Run Tasks for:
+Enhance your HashiCorp Cloud Platform Terraform (Terraform Cloud) workflows with AI-powered analysis using Amazon Bedrock. This module integrates seamlessly as a Run Task to provide:
 
-* Analyzing Terraform plan and generate short-summary
+* Intelligent Terraform plan analysis with concise, human-friendly summaries
+* Advanced function calling capabilities for extended analysis (e.g., AMI evaluation)
+* Responsible AI implementation with customizable guardrails
 
-* Function calling for other API-based analysis (e.g AMI analysis)
+### Key Features
 
-### Terraform plan summary
+1. **AI-Powered Plan Summaries**
+   * Generate clear, natural language summaries of Terraform plan outputs
+   * Quickly understand the impact of infrastructure changes
+   ![Example](./images/example.png)
 
-Summarize Terraform plan output in human friendly natural language.
-![Example](./images/example.png)
+2. **Extensible Analysis with Function Calling**
+   * Leverage AI to perform additional analyses, such as AMI evaluation
+   * Easily extend to other API-based tools and services
+   ![Example2](./images/example2.png)
 
-### Function calling (AMI analysis)
+3. **Responsible AI Integration**
+   * Implement Amazon Bedrock guardrails tailored to your organization's policies
+   * Ensure ethical and compliant AI usage in your infrastructure workflows
+   ![Example3](./images/example3.png)
 
-Use function calling to execute other tools, such as analyzing AMI data.
-![Example2](./images/example2.png)
+4. **Secure Architecture**
+   * Designed for deployment in a dedicated AWS account
+   * Optional AWS WAF integration for enhanced endpoint protection
+   * Adherence to AWS security best practices
 
-### Responsible AI
+5. **Seamless Terraform Cloud Integration**
+   * Operates as a native Run Task within your Terraform Cloud workflow
+   * Provides insights without disrupting existing processes
 
-Implement safeguards using Amazon Bedrock guardrails customized to your organization requirements and responsible AI policies
-![Example3](./images/example3.png)
+6. **Flexible and Customizable**
+   * Adapt the module to fit your specific organizational needs
+   * Easily configure and extend guardrails and analysis capabilities
 
 ## Architecture
 
 ![Diagram](./images/arch.png)
 
-This module must be deployed in AWS account with access to Amazon Bedrock model. We recommend you to use hub-spoke model, by deploying this module in one dedicated account.
-
-Please refer to the [best-practice](#best-practice) section below for more details.
+This module leverages a hub-spoke model, designed for deployment in a dedicated AWS account with Amazon Bedrock access. It utilizes AWS Lambda, CloudFront, and other AWS services to provide a scalable and secure solution.
 
 ## Prerequisites
 
-To use this module you need have the following:
+To implement this module, you'll need:
+
+1. An AWS account with appropriate credentials
+2. Access to Amazon Bedrock (default model: Claude 3 Sonnet)
+3. A HashiCorp Cloud Platform (HCP) Terraform account
+
+## Getting Started
+
+For detailed setup instructions and best practices, please refer to the sections below:
+
+* [Usage](#usage)
+* [Best Practices](#best-practice)
 
-1. AWS account and credentials
-1. Amazon Bedrock model access (default model is `Claude 3 Sonnet`)
-1. HCP Terraform account
+Enhance your Terraform workflows with AI-powered insights while maintaining security and responsible AI practices.
 
 ## Usage
 

VERSION

@@ -1 +1 @@
-v0.0.3
+v0.0.4

lambda/runtask_fulfillment/ai.py

@@ -8,6 +8,7 @@
 from runtask_utils import generate_runtask_result
 from tools.get_ami_releases import GetECSAmisReleases
 from utils import logger, stream_messages, tool_config
+import xml.etree.ElementTree as ET
 
 # Initialize model_id and region
 model_id = os.environ.get("BEDROCK_LLM_MODEL")
@@ -180,15 +181,27 @@ def eval(tf_plan_json):
         # Add response to message history
         messages.append(response)
 
-    result = response["content"][0]["text"]
+    # Try to parse output as XML and look for the <output> tag
+    try:
+        root = ET.fromstring(response["content"][0]["text"])
+        result = root.find("result").text
+        logger.info("Parsed : {}".format(result))
+    except Exception as e:
+        result = response["content"][0]["text"]
+        logger.info("Non Parsed : {}".format(result))
 
     #####################################################################
     ######### Third, generate short summary                     #########
     #####################################################################
 
     logger.info("##### Generating short summary #####")
     prompt = f"""
-        Can you provide a short summary with maximum of 150 characters of the infrastructure changes?
+        List the resources that will be created, modified or deleted in the following terraform plan using the following rules:
+        - Provide summary of the infrastructure changes
+        - Highlight major components of the changes such as what Terraform modules is executed
+        - Summarize what each Terraform modules does
+        - Highlight any resources that being replaced or deleted
+        - Highlight any outputs if available
 
         <terraform_plan>
         {tf_plan_json["resource_changes"]}
@@ -214,7 +227,7 @@ def eval(tf_plan_json):
 
     guardrail_status, guardrail_response = guardrail_inspection(str(description))
     if guardrail_status:
-        results.append(generate_runtask_result(outcome_id="Plan-Summary", description="Summary of Terraform plan", result=description[:700]))
+        results.append(generate_runtask_result(outcome_id="Plan-Summary", description="Summary of Terraform plan", result=description[:9000])) # body max limit of 10,000 chars
     else:
         results.append(generate_runtask_result(outcome_id="Plan-Summary", description="Summary of Terraform plan", result="Output omitted due to : {}".format(guardrail_response)))
         description = "Bedrock guardrail triggered : {}".format(guardrail_response)
@@ -225,7 +238,8 @@ def eval(tf_plan_json):
     else:
         results.append(generate_runtask_result(outcome_id="AMI-Summary", description="Summary of AMI changes", result="Output omitted due to : {}".format(guardrail_response)))
 
-    return description, results
+    runtask_high_level ="Terraform plan analyzer using Amazon Bedrock, expand the findings below to learn more. Click `view more details` to get the detailed logs"
+    return runtask_high_level, results
 
 def guardrail_inspection(input_text, input_mode = 'OUTPUT'):