Codex CLI v0.165 adds built-in Vertex AI provider and enables memories by default

Codex CLI v0.165.0 introduces a built-in Google Cloud Vertex AI provider configured via model_provider = "gcp" and turns on memory v2 by default for fresh profiles. The release also enables mid-conversation reasoning effort changes and closes a marketplace URL validation gap.

• Vertex AI is configured with model_provider = "gcp" in user-level config
• Memory v2 is on by default for fresh profiles, stored in ~/.codex/memories/
• Reasoning effort can now change mid-conversation on supported models
• A marketplace URL validation gap that could redirect Git to another host was closed