AWS automation AI agent

# **AWS AI Agent Project: Upload Local Files to Amazon S3** ## Project Goal Create an educational **AWS AI agent** that uploads files from a local Windows computer to an **Amazon S3 bucket**. The project must include two versions: 1. A standard AI agent using Python, Boto3, and **AWS services** 2. An agent using **AWS AgentCore** The main purpose is to show how an **AWS agent** can understand a request, perform an upload, read AWS results and logs, detect problems, and decide what to do next. ## Example Request The user may write: Upload all PDF files from my local Reports folder to my AWS S3 bucket. The agent must: 1. Find the requested local files. 2. Ask for confirmation. 3. Upload the files to **Amazon S3**. 4. Check that every file was copied correctly. 5. Read AWS errors and logs. 6. Retry temporary failures. 7. Stop and explain serious problems. 8. Produce a final report. ## File Verification For every file, the solution must: 1. Calculate a SHA-256 hash before upload. 2. Upload the file to **Amazon S3**. 3. Calculate or retrieve the hash of the uploaded S3 object. 4. Compare the local and S3 hashes. 5. Report success only when the hashes match. The agent must not report success when the copied file is different from the local file. ## Agent Thinking Scenarios The project must demonstrate the **AWS agent** handling several situations: * Successful S3 upload * Hash mismatch * Missing local file * Invalid S3 bucket * AWS IAM permission error * Interrupted upload * Some files succeed and some fail * File already exists in S3 * Temporary AWS network error * Ambiguous user request The agent must read AWS tool results and logs, then decide whether to: * Retry * Skip * Stop * Ask the user * Re-upload the file * Report success ## Required AWS Services The project should use: * **Amazon S3** * **AWS IAM** * **AWS SDK for Python using Boto3** * **AWS CloudWatch Logs**, where applicable * **Amazon Bedrock**, if used for the LLM * **AWS AgentCore** for the second version ## Required Deliverables The freelancer must provide: ### 1. Two Working AWS Projects * Standard AWS AI agent * AWS AgentCore agent ### 2. Complete Source Code The code must include: * Installation instructions * Configuration examples * AWS IAM policy * Error handling * Logging * Hash verification * Retry logic * Cleanup instructions ### 3. Step-by-Step Video The video must show: * AWS account setup * IAM setup * S3 bucket creation * Local installation * Standard agent setup * AWS AgentCore setup * File upload * Hash verification * AWS log review * Error handling * AWS resource cleanup ### 4. Word Document The Word document must include step-by-step instructions with screenshots for every major step. It must show: * AWS Console screens * IAM configuration * S3 configuration * AWS AgentCore configuration * Terminal commands * Successful upload * Hash comparison * Logs * Errors * Final verification ## Acceptance Criteria The project will be accepted only when: 1. I can install both projects in my own **AWS account**. 2. I can upload files from my local computer to **Amazon S3**. 3. The agent calculates a SHA-256 hash before upload. 4. The agent verifies the uploaded S3 file. 5. The local and S3 hashes are compared. 6. The agent reports success only when the hashes match. 7. The agent can read AWS errors and logs. 8. The agent retries temporary AWS failures. 9. The agent stops retrying when the problem is an IAM permission error. 10. The agent can recover when only some files fail. 11. The standard agent and AWS AgentCore versions perform the same test scenarios. 12. The video shows the complete process. 13. The Word document contains screenshots for every major AWS step. 14. I can reproduce the project without additional paid help. ## Important This must be a real **AWS agent project**, not only a simple S3 upload script. The agent must: * Understand the user request * Select AWS tools * Read AWS responses * Read logs * Detect problems * Decide the next action * Verify the final result Python and AWS SDK code should perform the actual upload and hash calculation. The AI agent should understand the task, control the workflow, interpret AWS results, and decide what happens next.