Feature Request: Add clear guidance on error handling patterns and recommended retry strategies

### Is your feature request related to a problem? Please describe.

## Overview
The `base/account-sdk` exposes various methods that may fail due to network issues, invalid input, or upstream RPC errors. However, the documentation does not clearly explain how developers are expected to handle these errors or which retry patterns are considered safe.

## Problem Description
When integrating the SDK, developers may be unsure about:
- which errors are transient vs permanent
- whether retries are safe for certain operations
- how to distinguish configuration errors from runtime failures
- what a recommended error-handling pattern looks like in practice

As a result, teams often implement inconsistent or unsafe retry logic.

### Describe the solution you'd like

### Proposed Solution
- Add a documentation section describing common error categories returned by the SDK.
- Explain which errors are retryable and which should fail fast.
- Provide simple example patterns for handling errors and retries safely.
- Keep guidance high-level and avoid prescribing strict application logic.

### Acceptance Criteria
- Documentation includes a clear error-handling and retry guidance section.
- Developers can easily identify safe retry scenarios.
- No changes are made to SDK behavior or public APIs.
- Guidance is concise, practical, and easy to maintain.

Start by reviewing the base/account-sdk methods and existing documentation to identify the error categories and operation types discussed in the issue. Add a concise section covering transient versus permanent errors, safe retry scenarios, configuration failures, and practical handling patterns; done means the guidance is clear and SDK behavior and public APIs are unchanged.

Written by the indexing model from the issue text.