Bug Description

When all auth profiles for a provider timeout (possible rate limit), Clawdbot does not fall back to the next model in agents.defaults.model.fallbacks. Instead, the request is aborted.

Expected Behavior

When google-antigravity:[email protected] times out and there are no other profiles for that provider, Clawdbot should fall back to openai-codex/gpt-5.2 (configured in agents.defaults.model.fallbacks).

Actual Behavior

The request is aborted with FailoverError: Request was aborted and no fallback is attempted.

Logs

[agent/embedded] Profile google-antigravity:[email protected] timed out (possible rate limit). Trying next account...
Embedded agent failed before reply: Request was aborted
[agent/embedded] Profile google-antigravity:[email protected] timed out (possible rate limit). Trying next account...
Embedded agent failed before reply: Request was aborted

This pattern repeated multiple times during heartbeat cron runs overnight.

Config

{
  "agents": {
    "defaults": {
      "model": {
        "primary": "google-antigravity/claude-opus-4-5-thinking",
        "fallbacks": ["openai-codex/gpt-5.2"]
      }
    }
  }
}

Environment

• Clawdbot version: 2026.1.16-2
• OS: macOS (Darwin 25.2.0 arm64)
• Node: v25.2.1

Notes

According to the docs:

If all profiles for a provider fail, Clawdbot moves to the next model in agents.defaults.model.fallbacks.

This should include timeouts that exhaust profile rotation.