Add documentation for GitHub Action workflow

Co-authored-by: guygregory <16044916+guygregory@users.noreply.github.com>

Author: Copilot

.github/workflows/README.md

@@ -0,0 +1,72 @@
+# Azure OpenAI Responses API Test Workflow
+
+This GitHub Action tests the `responses-basic-aoai-v1.py` script to ensure it returns a valid response from Azure OpenAI.
+
+## How to Run
+
+1. Go to the **Actions** tab in the GitHub repository
+2. Select **"Test Azure OpenAI Responses API"** workflow
+3. Click **"Run workflow"** button
+4. Click **"Run workflow"** to confirm
+
+## Required Environment Secrets
+
+The workflow uses the following secrets from the `responses` environment:
+
+- `AZURE_OPENAI_API_KEY` - Your Azure OpenAI API key
+- `AZURE_OPENAI_V1_API_ENDPOINT` - Your Azure OpenAI v1 API endpoint (e.g., `https://your-resource.openai.azure.com/openai/v1/`)
+- `AZURE_OPENAI_API_MODEL` - The model name to use (e.g., `gpt-4o`)
+
+## Test Results
+
+The workflow generates test artifacts that include:
+
+### JSON Results (`test-results.json`)
+```json
+{
+  "test_last_run_date": "2025-07-20T23:01:29Z",
+  "output": "Why don't scientists trust atoms? Because they make up everything!",
+  "pass_fail": "PASS",
+  "error_code": ""
+}
+```
+
+### Human-Readable Summary (`test-summary.txt`)
+```
+Azure OpenAI Responses API Test Results
+========================================
+Test Run Date: 2025-07-20T23:01:29Z
+Result: PASS
+Error Code: 
+
+Output:
+Why don't scientists trust atoms? Because they make up everything!
+```
+
+## Test Criteria
+
+The test passes if:
+- The Python script executes without errors
+- The script produces output
+- The output contains valid string content (not empty, no error indicators)
+
+The test fails if:
+- Environment variables are missing
+- The script fails to execute
+- No output is produced
+- Output contains error indicators (error, exception, traceback, failed, none, null)
+
+## Artifacts
+
+Test results are uploaded as artifacts with:
+- **Name**: `azure-openai-test-results`
+- **Retention**: 30 days
+- **Contents**: Both JSON and human-readable results
+
+## Troubleshooting
+
+If the workflow fails:
+1. Check that all required environment secrets are set in the `responses` environment
+2. Verify that the Azure OpenAI service is accessible
+3. Review the workflow logs for specific error messages
+4. Check the uploaded artifacts for detailed test results