BEFORE — ccae953c46 Environment: Linux; real T3 server; fresh isolated data; Codex provider. Entry point: source CLI (node apps/server/src/bin.ts). reasoning effort $ t3 thread start . "Reply cli ok" --provider codex --model gpt-6.1-sol --option reasoningEffort=high --wait --json ERROR Unrecognized flag: --option in command t3 thread start Exit: 1 boolean false $ t3 thread start . "Reply cli ok" --provider codex --model gpt-6.1-sol --option reasoningEffort=high --option fastMode=false --wait --json ERROR Unrecognized flag: --option in command t3 thread start Unrecognized flag: --option in command t3 thread start Exit: 1 duplicate option $ t3 thread start . "Reply cli ok" --provider codex --model gpt-6.1-sol --option reasoningEffort=high --option reasoningEffort=low --wait --json ERROR Unrecognized flag: --option in command t3 thread start Unrecognized flag: --option in command t3 thread start Exit: 1 missing provider and model $ t3 thread start . "Reply cli ok" --option reasoningEffort=high --wait --json ERROR Unrecognized flag: --option in command t3 thread start Exit: 1