Short answer: Add
@once-agent/mcp to your Cursor project so coding-agent workflows can evaluate and protect external writes that may otherwise be repeated after an ambiguous outcome.Why this matters
AI coding agents can create software that performs payments, refunds, bookings, orders, provisioning, deployments, messages and other externally visible writes.
If one of those operations succeeds but its response is lost, another execution can repeat the same real-world effect.
Add Once through MCP
npx -y @once-agent/mcpOnce is published as an npm MCP server and registered under io.github.stringsofthemind-oss/once.
Cursor project configuration
{
"mcpServers": {
"once": {
"type": "stdio",
"command": "npx",
"args": ["-y", "@once-agent/mcp"]
}
}
}The Once repository includes a ready-to-copy .cursor/mcp.json project configuration.
View the Cursor + Once MCP example on GitHub
What the Once MCP server exposes
- project assessment for likely consequential writes
- retry-risk explanations
- setup and protection planning
- approved protection application
- Once Cloud connection verification
- live proof and current claim boundaries
Claim boundary: Once does not claim universal exactly-once execution. Safety depends on stable operation identity, durable Once state, the provider integration and sufficiently authoritative provider truth. When an outcome cannot be established safely, Once can preserve uncertainty instead of blindly repeating the action.