Codex provider fails to parse 'prolite' plan type

[yossy6028]

Description

Codex provider fails with a decode error when the OpenAI API returns prolite as the plan_type. This is a new plan tier (Pro Lite) that is not yet included in the expected enum variants.

Error

Error: Codex connection failed: failed to fetch codex rate limits: Decode error for
https://chatgpt.com/backend-api/wham/usage: unknown variant `prolite`, expected one of
`guest`, `free`, `go`, `plus`, `pro`, `free_workspace`, `team`, `business`, `education`,
`quorum`, `k12`, `enterprise`, `edu` at line 5 column 24

API Response (relevant fields)

{
  "plan_type": "prolite",
  "rate_limit": {
    "allowed": true,
    "limit_reached": false,
    "primary_window": {
      "used_percent": 14,
      "limit_window_seconds": 18000
    },
    "secondary_window": {
      "used_percent": 19,
      "limit_window_seconds": 604800
    }
  },
  "additional_rate_limits": [
    {
      "limit_name": "GPT-5.3-Codex-Spark",
      "metered_feature": "codex_bengalfox"
    }
  ]
}

The usage data is returned correctly — it's just the plan type enum that needs prolite added.

Environment

• CodexBar: v0.20
• macOS: 15 (Sequoia)
• Architecture: Apple Silicon
• Codex CLI: v0.117.0

Expected Behavior

CodexBar should recognize prolite as a valid plan type and display usage data normally.