itzzjb/DockAI - AI Dockerfile Generation Framework
The Customizable AI Framework that generates, validates, and optimizes production-ready Dockerfiles.
View on GitHubTrust Signals
- Scorecard Score
- not yet scored
- Maintenance Recency
- Activelast commit Apr 1, 2026
- License
- MIT
Pinned Snippet
uses: itzzjb/dockai@12a147d0460abe40bd15d00068421cc97e61e153 # v4.0.7tags can be moved; commit SHAs can't. why a SHA?
Inputs
| name | description | required | default |
|---|---|---|---|
| llm_provider | LLM provider to use (openai, azure, gemini, anthropic, ollama) | no | openai |
| openai_api_key | OpenAI API Key (required for openai provider) | no | — |
| azure_openai_api_key | Azure OpenAI API Key (required for azure provider) | no | — |
| azure_openai_endpoint | Azure OpenAI endpoint URL (required for azure provider) | no | — |
| azure_openai_api_version | Azure OpenAI API version | no | 2024-02-15-preview |
| google_api_key | Google API Key (required for gemini provider) | no | — |
| anthropic_api_key | Anthropic API Key (required for anthropic provider) | no | — |
| project_path | Path to the project to dockerize | no | . |
| model_analyzer | Model for analyzer agent (default: gpt-4o-mini for openai, gemini-1.5-flash for gemini) | no | "" |
| model_blueprint | Model for blueprint agent (Chief Architect) | no | "" |
| model_generator | Model for generator agent (default: gpt-4o for openai, gemini-1.5-pro for gemini) | no | "" |
| model_reviewer | Model for security reviewer agent | no | "" |
| model_reflector | Model for failure reflection agent | no | "" |
| model_error_analyzer | Model for error classification agent | no | "" |
| model_iterative_improver | Model for iterative improvement agent | no | "" |
| max_retries | Max attempts to fix a failing Dockerfile | no | 3 |
| skip_hadolint | Skip Hadolint Dockerfile linting (true/false) | no | false |
| skip_security_scan | Skip Trivy security scan (true/false) | no | false |
| strict_security | Fail on any vulnerability (true/false) | no | false |
| max_image_size_mb | Max image size in MB (0 to disable) | no | 500 |
| skip_health_check | Skip health checks during validation (true/false) | no | false |
| skip_security_review | Skip AI security review for faster generation (auto-skipped for scripts) | no | false |
| validation_memory | Memory limit for container validation (e.g., 512m, 1g) | no | 512m |
| validation_cpus | CPU limit for container validation | no | 1.0 |
| validation_pids | Max processes for container validation | no | 100 |
| max_file_chars | Max characters to read per file (default: 200000) | no | 200000 |
| max_file_lines | Max lines to read per file (default: 5000) | no | 5000 |
| truncation_enabled | Force-enable smart truncation for critical files (true/false) | no | false |
| token_limit | Token threshold that auto-enables truncation (default: 50000) | no | 50000 |
| use_rag | Enable RAG (Retrieval-Augmented Generation) for intelligent context retrieval (true/false, default: true in v4.0) | no | true |
| embedding_model | HuggingFace embedding model for RAG (default: all-MiniLM-L6-v2) | no | all-MiniLM-L6-v2 |
| read_all_files | Read all source files for complete context (true/false, default: true) | no | true |
| analyzer_instructions | Custom instructions for the analyzer (project discovery agent) | no | "" |
| generator_instructions | Custom instructions for the generator (Dockerfile creation agent) | no | "" |
| blueprint_instructions | Custom instructions for the blueprint agent (Chief Architect) | no | "" |
| reviewer_instructions | Custom instructions for security review (auditor agent) | no | "" |
| reflector_instructions | Custom instructions for failure analysis (detective agent) | no | "" |
| error_analyzer_instructions | Custom instructions for error classification (troubleshooter agent) | no | "" |
| iterative_improver_instructions | Custom instructions for applying fixes (surgical fix agent) | no | "" |
| prompt_analyzer | Custom prompt - project discovery and deduction agent | no | "" |
| prompt_blueprint | Custom prompt - blueprint agent (Chief Architect) | no | "" |
| prompt_generator | Custom prompt - Dockerfile reasoning and creation agent | no | "" |
| prompt_reviewer | Custom prompt - security auditor and threat analysis agent | no | "" |
| prompt_reflector | Custom prompt - detective and post-mortem analysis agent | no | "" |
| prompt_error_analyzer | Custom prompt - troubleshooter and error classification agent | no | "" |
| prompt_iterative_improver | Custom prompt - surgeon for precise corrections agent | no | "" |
| llm_caching | Enable in-process LLM response caching (true/false) | no | true |
| ollama_base_url | Base URL for Ollama when using llm_provider=ollama | no | http://localhost:11434 |
| google_cloud_project | Google Cloud project ID for Gemini (optional) | no | "" |
| collect_metrics | Enable metrics collection (true/false) | no | false |
| enable_tracing | Enable OpenTelemetry tracing (true/false) | no | false |
| tracing_exporter | Tracing exporter type (console, otlp) | no | console |
| otlp_endpoint | OTLP endpoint URL for trace export | no | http://localhost:4317 |
| otel_service_name | Service name for OpenTelemetry traces | no | dockai |
| langchain_tracing_v2 | Enable LangSmith/LangChain tracing (true/false) | no | false |
| langchain_api_key | LangSmith API key (required when langchain_tracing_v2=true) | no | — |
| langchain_project | LangSmith project name | no | dockai |
| langchain_endpoint | LangSmith endpoint URL | no | https://api.smith.langchain.com |
Outputs
no outputs