Update to Codex Process Jobs

{
  "name": "rerun",
  "description": "Launch a finished Codex Process Jobs record again as a new detached job using its validated persisted argv, working directory, and execution mode. Use only when the user explicitly asks or approves rerunning a specific completed, failed, or cancelled build, test, benchmark, inference run, data job, or repair.",
  "included_files": [
    {
      "relative_path": "agents/openai.yaml",
      "size_in_bytes": 180
    }
  ],
  "skill_md_contents": "---\nname: rerun\ndescription: Launch a finished Codex Process Jobs record again as a new detached job using its validated persisted argv, working directory, and execution mode. Use only when the user explicitly asks or approves rerunning a specific completed, failed, or cancelled build, test, benchmark, inference run, data job, or repair.\n---\n\n# Rerun Process Job\n\nResolve `<plugin-root>` two directories above this file and run:\n\n```text\nnode \"<plugin-root>/scripts/job.mjs\" rerun <job-id> [options] --json\n```\n\nNever search memory for CPJ work; use validated CPJ state.\n\nThe user-visible parent owns every CPJ rerun and completion. Never delegate\nlocal process execution, launch, waiting, monitoring, or CPJ ownership to a\nspawned subagent. A subagent can analyze independent material only.\n\nRequire a specific source job. If identity is unclear, use `$status` once to\nfind it. Never reconstruct argv from displayed command text, logs, or model\nmemory. The controller reads the private validated record, refuses active jobs,\nand creates a fresh job ID and logs with `rerunOf` lineage.\n\nA rerun repeats the invocation, not the historical environment: files,\ndependencies, environment variables, credentials, devices, and external state\nmay have changed. Mention a material change before launch; do not claim exact\nreproducibility.\n\nNever rerun automatically from a completion notice. A direct user request\nauthorizes an ordinary non-critical rerun. For a `CRITICAL` job, explain that\nthe command may repeat repair, migration, firmware, or destructive effects and\nobtain explicit risk-aware approval immediately before adding `--force`.\n\nAdd `--goal-mode` only for an explicitly active Goal. Optional flags are\n`--no-notify`, `--notify-user`, and `--no-notify-user`. If the durable state\ndirectory is not writable in the current sandbox, request the same narrow\ncontroller escalation described by `$start`.\n\nTreat a successful rerun as the same hard release boundary as `$start`: report\nthe new job ID and its source job ID in no more than two short sentences. Say\nthat a completion notification should appear when it finishes, then end the\nturn without status, tail, result, wait, sleep, or process monitoring. Do not\nnarrate controller mechanics, persisted metadata, argv, cwd, or validation\nunless the user explicitly asks.\nIf the controller refuses because the source remains active, its working\ndirectory disappeared, or legacy shell semantics cannot be preserved, report\nthat refusal and do not improvise a replacement command.\n"
}