emisar
View on GitHubLet AI agents investigate and operate infrastructure through declared actions, bounded by policy and host-side checks. Approvals when required, with an audit trail. By Protectorate.
Emisar provides an MCP interface for AI agents to operate infrastructure through declared, policy-bounded actions. A host runner validates actions locally; approvals, redaction, and audit trails help control production access.
Use Cases
Let AI agents run approved infrastructure actionsManage Linux hosts through declared MCP actionsRequire human approval for risky operationsAudit agent-requested infrastructure changesCreate and distribute reusable infrastructure action packs
Built With
- Language
- Elixir
- Frameworks
- Phoenix
Tags
MCP server · AI agents · infrastructure · DevOps · policy-as-code · security · action packs · human approval · audit trail · self-hosted