# Ethics Vibe Check — environment

# Second-opinion provider: gemini (default) | ollama | mock
# mock is for tests/demos only
ETHICS_PROVIDER=gemini

# Optional model override for the active provider
# ETHICS_MODEL=gemini-2.0-flash
# ETHICS_MODEL=llama3.2

# Gemini (required when ETHICS_PROVIDER=gemini)
# https://aistudio.google.com/app/apikey
GEMINI_API_KEY=your_actual_gemini_api_key_here

# Ollama (when ETHICS_PROVIDER=ollama)
# OLLAMA_BASE_URL=http://127.0.0.1:11434
# OLLAMA_MODEL=llama3.2

# Optional: override where session drift history and learned concerns are stored.
# Defaults to ~/.ethics-vibe-check/data
# ETHICS_DATA_DIR=/absolute/path/to/your/ethics-data

# Optional: force a known session id (otherwise one is generated per MCP process)
# ETHICS_SESSION_ID=my-chat-42

# Optional: where `npm run install-skill` copies the companion skill
# CURSOR_SKILLS_DIR=/home/you/.cursor/skills