feat(codex): start a codex session at a chosen reasoning effort

[irisitymichaelgrundberg]

Part of #513 (the Codex half).

codexConfig takes a new reasoningEffort, and a codex session starts with --config model_reasoning_effort=<level>.

Why

A caller can pick codex's model per session through codexConfig.model, but not its reasoning effort, so every session starts at whatever ~/.codex/config.toml says. Claude sessions already take an effort per session through effort.

What it does

• CODEX_REASONING_EFFORTS in src/types/session.ts lists the levels codex accepts as of codex-cli 0.154.0: none, minimal, low, medium, high, xhigh, max and ultra. Which of them a given model honours is codex's call. CodexConfig gets a matching reasoningEffort field.
• The schema admits only those words. CodexConfigSchema validates reasoningEffort as an enum built from that list, so both POST /api/sessions and POST /api/quick-start reject anything else.
• The registry emits one literal per level. An argv token can't splice a parameter into a literal, and codex wants model_reasoning_effort=<level> as one --config value. The Codex entry therefore declares a reasoningEffort enum parameter and one --config model_reasoning_effort=<level> argument per level, each gated on its own value. A level the enum doesn't know emits nothing. The registry needed no new argument shape for this.
• The arguments come after --model and before the resume <id> subcommand, so a resumed session keeps its effort too.

The "Codex specifics" list in docs/architecture-invariants.md now names the two --config values codex takes from codexConfig. codexConfig already round-trips through SessionState, so the effort survives a respawn and a Codeman restart. npm run generate:cli-catalog -- --check reports the catalogue in sync, because launch arguments aren't part of it.

Testing

• npm test passes, as do typecheck, lint, format and the catalogue check.
• test/cli-registry-spawn-golden.test.ts gains two cases:
  • one renders the launch command for every level;
  • one pins the order with a model and a resume id, codex --model gpt-5 --config model_reasoning_effort=high resume roll_42.
• The new test/codex-reasoning-effort-schema.test.ts checks that both create schemas accept every level, and that both reject unknown words and shell characters.
• test/routes/external-cli-bypass-clamp.test.ts checks that a non-granted owner in multi-user mode keeps the effort when the clamp forces bypass off.
• By hand, codex --config model_reasoning_effort=xhigh starts codex-cli 0.154.0 with gpt-5.6-sol xhigh in its header, where the user config says medium. Codex needs no extra quoting around the level.

No changeset, following the repository's convention for contributors.

🤖 Generated with Claude Code