When multiple configuration keys exist for related settings (e.g., “deployment vs model identity”), verify the runtime meaning of each key and ensure examples/placeholders reflect the correct mapping. Also ensure any conditional config logic (feature flags/version gating) is accurate, necessary, and easy to read, and that expression comparisons use the correct types (boolean vs string). Checklist: - Identify what each config key controls (routing/deployment vs canonical model identity) and avoid swapped assignments. - Keep placeholders aligned with the corrected semantics, and cross-check any fallback/resolution keys (e.g., a CLI’s primary model identifier). - Prefer simplifying templated conditions when the output is guaranteed (remove unnecessary branching). - In expression/config logic, compare booleans to booleans (not `'true'`), and ensure you only use supported contexts/variables. Example (environment variables): ```shell # Correct mapping: deployment name routes the request; model identity drives model capabilities. export COPILOT_PROVIDER_BASE_URL=https://YOUR-RESOURCE-NAME.openai.azure.com export COPILOT_PROVIDER_TYPE=azure export COPILOT_PROVIDER_API_KEY=YOUR-AZURE-API-KEY export COPILOT_PROVIDER_AZURE_API_VERSION=YOUR-AZURE-API-VERSION export COPILOT_PROVIDER_MODEL_ID=YOUR-MODEL-NAME export COPILOT_PROVIDER_WIRE_MODEL=YOUR-DEPLOYMENT-NAME # Optionally confirm whether COPILOT_MODEL should match which identifier the CLI resolves against. ```