# ---------------------------------------------------------------------------
# Required
# ---------------------------------------------------------------------------
OPENAI_API_KEY=sk-***
# Optional, when enabling Claude Code:
# ANTHROPIC_API_KEY=
# For Claude Code OAuth instead, set:
# CLAUDE_CODE_OAUTH_TOKEN=

# ---------------------------------------------------------------------------
# JWT verification for AgentOS. Generate your JWT_VERIFICATION_KEY at os.agno.com.
# In production, the app refuses to serve traffic without a JWT_VERIFICATION_KEY or JWT_JWKS_FILE.
# ---------------------------------------------------------------------------
# Quote the value so parsers read the multi-line PEM as one variable:
# JWT_VERIFICATION_KEY="-----BEGIN PUBLIC KEY-----
# MIIBIjANBgkq...
# -----END PUBLIC KEY-----"
# JWT_JWKS_FILE=/app/jwks.json

# ---------------------------------------------------------------------------
# AgentOS URL used by the scheduler and MCP server.
# Set this to the public URL of your AgentOS instance.
# ---------------------------------------------------------------------------
# AGENTOS_URL=http://127.0.0.1:8000

# ---------------------------------------------------------------------------
# MCP OAuth — set MCP_CONNECT_SECRET to turn the /mcp endpoint into its own
# OAuth 2.1 authorization server, which is how claude.ai and ChatGPT (web) connect.
# Connecting to the MCP server asks for this secret on a consent page.
# Generate a strong one (≥16 chars):
#   openssl rand -base64 32
# ---------------------------------------------------------------------------
# MCP_CONNECT_SECRET=

# ---------------------------------------------------------------------------
# Web access uses Parallel SDK when authenticated,
# and falls back to Parallel's keyless MCP server for local development.
# ---------------------------------------------------------------------------
# PARALLEL_API_KEY=

# ---------------------------------------------------------------------------
# Slack — set both to enable the Slack interface.
#   SLACK_BOT_TOKEN       Bot User OAuth Token (xoxb-***)
#   SLACK_SIGNING_SECRET  Signing Secret (Basic Information → App Credentials)
# ---------------------------------------------------------------------------
# SLACK_BOT_TOKEN=
# SLACK_SIGNING_SECRET=

# ---------------------------------------------------------------------------
# Database
# ---------------------------------------------------------------------------
# DB_HOST=localhost
# DB_PORT=5432
# DB_USER=ai
# DB_PASS=ai
# DB_DATABASE=ai
# DB_DRIVER=postgresql+psycopg

# ---------------------------------------------------------------------------
# Optional GitHub access for coding agents (forwarded to sandboxes).
# Scope the token to the repositories the agents should use.
# ---------------------------------------------------------------------------
# GH_TOKEN=
# GITHUB_READ_TOKEN=             # Optional separate private-checkout token; defaults to GH_TOKEN.
# Commits default to AgentOS <agentos@localhost>. Override the author/committer:
# GIT_AUTHOR_NAME=
# GIT_AUTHOR_EMAIL=
# GIT_COMMITTER_NAME=
# GIT_COMMITTER_EMAIL=

# ---------------------------------------------------------------------------
# Optional sandbox — enable SANDBOX_ENABLED in app/sandbox.py; see docs/harnesses.md
# ---------------------------------------------------------------------------
# Docker locally; set agentcore on your API host.
# SANDBOX_PROVIDER=docker

# Local setup: scripts/aws.py reads these settings from .env.
# AWS_PROFILE=your-profile
# AWS_REGION=us-east-1
# DEPLOYMENT_NAME=agentos-sandbox
# BUCKET_NAME=your-checkpoint-bucket  # Omit to derive a name.
# CHECKPOINT_PREFIX=agentos/checkpoints/

# After setup: add the printed values to your API host's environment.
# AGENTCORE_RUNTIME_ARN=
# AGENTCORE_CHECKPOINT_BUCKET=
# AGENTCORE_CHECKPOINT_PREFIX=agentos/checkpoints/

# Deployment overrides (see README.md)
# AZURE_RESOURCE_GROUP=agentos
# AZURE_LOCATION=eastus