| prompt | The prompt to send to the AI model | yes | — |
| openai_api_key | OpenAI API key | no | — |
| anthropic_api_key | Anthropic API key | no | — |
| google_ai_api_key | Google AI Studio API key | no | — |
| azure_openai_api_key | Azure OpenAI API key | no | — |
| azure_openai_endpoint | Azure OpenAI endpoint URL | no | — |
| azure_openai_deployment | Azure OpenAI deployment name | no | — |
| mistral_api_key | Mistral AI API key | no | — |
| huggingface_api_key | Hugging Face API key | no | — |
| openrouter_api_key | OpenRouter API key | no | — |
| litellm_api_key | LiteLLM API key | no | — |
| litellm_base_url | LiteLLM base URL | no | — |
| openai_compatible_api_key | OpenAI-compatible API key | no | — |
| openai_compatible_base_url | OpenAI-compatible base URL | no | — |
| aws_access_key_id | AWS Access Key ID for Bedrock/SageMaker | no | — |
| aws_secret_access_key | AWS Secret Access Key | no | — |
| aws_region | AWS Region | no | us-east-1 |
| aws_session_token | AWS Session Token | no | — |
| bedrock_model_id | AWS Bedrock model ID | no | — |
| sagemaker_endpoint | Amazon SageMaker endpoint | no | — |
| google_vertex_project | Google Cloud project ID for Vertex AI | no | — |
| google_vertex_location | Google Cloud location | no | us-central1 |
| google_application_credentials | GCP service account JSON (base64 encoded) | no | — |
| provider | AI provider (openai, anthropic, google-ai, vertex, azure, bedrock, mistral, huggingface, openrouter, litellm, ollama, sagemaker, openai-compatible) | no | auto |
| model | Specific model to use | no | — |
| temperature | Sampling temperature (0.0-2.0) | no | 0.7 |
| max_tokens | Maximum tokens in response | no | 4096 |
| system_prompt | System prompt for context | no | — |
| command | CLI command: generate, stream, batch | no | generate |
| image_paths | Comma-separated image paths | no | — |
| pdf_paths | Comma-separated PDF paths | no | — |
| csv_paths | Comma-separated CSV paths | no | — |
| video_paths | Comma-separated video paths | no | — |
| thinking_enabled | Enable extended thinking | no | false |
| thinking_level | Thinking level: minimal, low, medium, high | no | medium |
| thinking_budget | Thinking token budget | no | 10000 |
| enable_analytics | Enable usage analytics and cost tracking | no | false |
| enable_evaluation | Enable response quality evaluation | no | false |
| output_format | Output format: text, json | no | text |
| output_file | Output file path | no | — |
| enable_tools | Enable MCP tools | no | false |
| mcp_config_path | Path to .mcp-config.json file | no | — |
| post_comment | Post AI response as PR/issue comment | no | false |
| update_existing_comment | Update existing NeuroLink comment instead of creating new | no | true |
| comment_tag | HTML comment tag to identify NeuroLink comments | no | neurolink-action |
| github_token | GitHub token for PR/issue operations | no | ${{ github.token }} |
| timeout | Request timeout in seconds | no | 300 |
| debug | Enable debug logging | no | false |
| neurolink_version | NeuroLink CLI version to install | no | latest |
| working_directory | Working directory for CLI execution | no | . |