Opsy – Agent-first infrastructure management

Name: Opsy – Agent-first infrastructure management
Availability: InStock
Author: d36ugger

by d36ugger·Mar 10, 2026·1 point·0 comments

Visit Project View on HN

AI Analysis

●●SolidSolve My ProblemShip It

MCP integration lets agents draft Pulumi changes without ever touching prod credentials.

Strengths

•MCP integration lets agents like Claude Code natively propose infrastructure changes safely.
•Draft-and-approve workflow separates agent planning from execution, preventing potential accidental production destruction.
•CLI supports JSON output and quiet mode for easy automation scripting integration.

Weaknesses

•Relies on Pulumi underneath, limiting adoption for existing Terraform or CloudFormation shops.
•Approval workflow is manual; lacks any built-in policy-as-code automation for low-risk changes.

Post Description

Built this because AI agents keep getting access to infrastructure with zero guardrails. Opsy sits between the agent and your cloud — agents propose changes, you approve before anything runs. No credentials exposed to agents. They talk to Opsy through CLI or MCP, write drafts in YAML, and nothing touches prod without approval. Works with Claude Code or any agent that supports MCP out of the box.