{
  "id": 4797628,
  "title": "One Go binary between your engineers and production: how we broker SSH, kubectl, psql and RDP — and what it costs",
  "url": "https://urgent.news/2026/09/01/one-go-binary-between-your-engineers-and-production-how-we-broker-ssh",
  "topic": "tech",
  "section": "Tech",
  "published": "2026-09-01T05:38:27.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/alektoreef/one-go-binary-between-your-engineers-and-production-how-we-broker-ssh-kubectl-psql-and-rdp-and-563i"
  },
  "original_language": "en",
  "account": "One Go binary and PostgreSQL 16+ form the core of an access control system that manages SSH, kubectl, psql, and RDP connections. This system does not store or transmit secrets within the target systems; instead, it establishes connections from the controller to the target systems and securely injects the credentials during the session. By doing so, the system prevents the distribution of keys, kubeconfig, database passwords, and RDP credentials to individual machines, reducing the risk of unauthorized access spreading.\n\nThe controller verifies user permissions for each target and injects the relevant credentials into the connection. SSH, kubectl, and database access are all managed through this brokered approach, ensuring consistent security policies across all protocols. However, the system has limitations when it comes to SSH access. While Kubernetes and databases enforce permissions at the protocol level, SSH relies on a configurable command blocklist and does not provide the same level of protection. As a result, SSH access is treated more as a guardrail than a strict barrier, and it is recommended to use read-only accounts for enhanced security. This approach, while not perfect, offers a comprehensive solution to managing access across multiple systems while minimizing the risks associated with key distribution and unauthorized access.",
  "summary": "The contractor's public key is still in authorized_keys on eleven machines. You know because you put it there in March, for two weeks, until the migration was done. The contract ended in May. You remember the Ansible play that spread the keys. There was never a play that collected them back. This is not carelessness. It is what happens when access is handed out instead of granted . A key, a…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}