codex-cli 0.42.0 logs MCP startup at ERROR severity despite successful connections

What version of Codex is running?

codex-cli 0.42.0

Which model were you using?

gpt-5-codex

What platform is your computer?

Darwin 23.6.0 arm64 arm

What steps can reproduce the bug?

Summary: After upgrading to Codex CLI 0.42.0, every MCP server invoked through mcp-proxy reports a codex_core::mcp_connection_manager ERROR on startup even though the servers initialize and tools load correctly. VS Code (same MCP config) does not surface the message, so the regression appears isolated to the CLI.

Steps to Reproduce:
From the project root, run
RUST_LOG=codex_core::mcp_connection_manager=trace codex exec --cd /Users/xxx/Dev/xxx 'echo MCP test'
Observe CLI startup output.

What is the expected behavior?

Expected Result: MCP proxies negotiate 2025-06-18 and the CLI logs the connections at INFO/DEBUG without implying failure.

What do you see instead?

Actual Result: CLI emits ERROR entries like:
2025-10-02T01:43:04.463219Z ERROR codex_core::mcp_connection_manager: new mcp_servers: {...}
Immediately afterward, the CLI reports aggregated 57 tools from 6 servers, and subsequent tool use works.
Environment: macOS, Codex CLI 0.42.0, codex config using mcp-proxy 0.8.2, local MCP services behind Uvicorn plus remote Zapier/Context7 endpoints.

Additional information

Additional Notes: Manual proxy checks via ./.codex/mcp-proxy.sh --transport=streamablehttp … succeed; the regression only affects CLI logging. VS Code extension launching with the same MCP definitions produces no errors, so the new Rust client path in CLI likely needs log-level tuning.

Attachments: codex-cli-mcp-error-snippet.log (focused excerpt) and

codex-cli-mcp-error-snippet.log