Skip to content

CLI — AI auto mode

Auto mode fetches your repository's file tree, README, and CI/CD configuration from the hosting platform, then sends that context to an AI model to fill in the assessment answers automatically — no manual prompts required.

Supported AI providers

Provider Flag value API key env var Notes
OpenAI openai OPENAI_API_KEY GPT-4o and other OpenAI models
Anthropic anthropic ANTHROPIC_API_KEY Claude models
Google Gemini gemini GEMINI_API_KEY Gemini models
Ollama ollama — Local models, no API key needed

Basic usage

# OpenAI — provider auto-detected from the git remote URL
OPENAI_API_KEY=sk-... dm assess --auto --ai openai

# Specify a model explicitly
OPENAI_API_KEY=sk-... dm assess --auto --ai openai --model gpt-4o

# Anthropic
ANTHROPIC_API_KEY=... dm assess --auto --ai anthropic --model claude-3-5-sonnet-20241022

# Google Gemini
GEMINI_API_KEY=... dm assess --auto --ai gemini

# Ollama (runs locally, no API key needed)
dm assess --auto --ai ollama --model llama3

# Private repository — pass a token so the tool can read repo context
GITHUB_TOKEN=ghp_... dm assess --auto --ai openai

How it works

  1. The tool detects the git provider (GitHub, GitLab, or Bitbucket) from the origin remote URL.
  2. It fetches the repository file tree, README, and CI/CD configuration files using the provider's API.
  3. The repository context is sent to the chosen AI model alongside the list of criteria.
  4. The AI returns a JSON response with a true/false answer and a brief rationale for each criterion.
  5. Results are saved and displayed in the same format as interactive mode.

Note

The --provider flag is auto-detected from the origin remote URL of the current git repository. Public repositories do not require a token.

Warning

AI auto mode sends repository context to the chosen provider unless you use a local Ollama server. Only use it for repositories you are allowed to share with that provider.

All flags

See CLI flags reference for the full list of options.