{
  "slug": "railway-code-launches-cloud-agent-sessions-in-the-tui-a74d2fad2fca",
  "title": "railway code launches cloud-agent sessions in the TUI",
  "tools": [
    "railway"
  ],
  "tags": [
    "agent"
  ],
  "introduced_at": "2026-08-13T00:42:10.000Z",
  "announced_at": "2026-08-13T00:42:10.000Z",
  "effective_at": null,
  "date_basis": "source-publication",
  "status": "unknown",
  "how": {
    "incantation": "When launching repository work with a saved default harness, run `railway code`; to choose a harness, run `railway code --<harness>`. In v5.38.0, this opens the `railway ca` manage screen with the tree collapsed and the session starting in a full-width pane.",
    "examples": [
      "railway code",
      "railway code --<harness>"
    ],
    "gotchas": [
      "Version: v5.38.0",
      "--<harness>: Overrides the saved default harness; the release note does not enumerate accepted harness names.",
      "⌥f: Restores the tree in the manage screen.",
      "⌥n: Starts another session.",
      "Target resolution remains flags, linked directory, default project, then inline `railway ca setup`.",
      "The direct non-TUI path remains for `--rm`, `-- <args>` passthrough, piped or non-terminal runs, and `railway ca start`.",
      "Launch arguments including `--name`, `--variable`, `--env-file`, and `--refresh-auth` survive the TUI flow."
    ]
  },
  "why_it_matters": "When launching a saved or selected harness for repository work, open the cloud-agent manage screen with the session running in its pane.",
  "before": "`railway code` ssh-ed directly into the VM instead of opening the `railway ca` manage screen.",
  "sources": [
    "https://github.com/railwayapp/cli/releases/tag/v5.38.0",
    "https://github.com/railwayapp/cli/pull/1087"
  ],
  "rank": 0.92039,
  "judge": {
    "publish": true,
    "judge_model": "codex-cli/gpt-6-luna+typesafe/jev-1.13.0",
    "prompt_version": "extract-v7+enrich-v2+judge-v2x3+jev-publish-v1"
  },
  "review_state": "approved",
  "provenance": {
    "run_id": "wave1b-20261016",
    "verdict_id": "wave1b-20261016:wave1b-20261016-ac4301cf7f34bf1b-3:gate",
    "evidence_hash": "52e053a8370e05a7166fe051bf18938a110958eab02f09644a100499d56c3ed3"
  },
  "review": {
    "draft_hash": "4de7f97ea0eee6e99392bfa2f96c42d4e96dc9d9b7fcd261af77d08e0a566ff7",
    "reviewer": "publication-gate",
    "reviewed_at": "2026-10-07T00:00:00.000Z",
    "decision": "wave1b-20261016:gate",
    "reason": "3 of 3 judge runs approved with p_yes >= 0.9 (mean 0.92 >= 0.9; 2 needed) and Jev gave 0.74 (>= 0.6 needed)."
  }
}