feat: add session thinking effort flag

[codeErrorSleep]

Related Issue

Resolves #2535

Problem

Kimi Code users could not override the thinking effort when starting or resuming a session without changing
config.toml. This made it inconvenient to select different effort levels for interactive use, one-shot prompts,
scripts, and CI workflows.

What changed

• Added --effort <effort> support to the interactive TUI, --prompt mode, resumed sessions, and the v2 print path.
• Applied the override when creating a session and when resuming an existing session.
• Applied --model before --effort so agent-core validates the effort against the target model.
• Accepted any non-empty effort string at the CLI boundary, leaving model-specific validation and normalization to
  agent-core.
• Kept the override session-scoped without modifying config.toml.
• Synchronized the TUI from the resulting session state so normalized effort values are displayed correctly.
• Added coverage for argument parsing, new and resumed sessions, model/effort ordering, v1 prompt mode, v2 print
  mode, and TUI startup.
• Updated the English and Chinese CLI documentation and added a patch changeset.

Validation completed:

• Targeted Vitest suite: 197 tests passed.
• pnpm --filter @moonshot-ai/kimi-code typecheck
• pnpm --filter @moonshot-ai/kimi-code build
• pnpm lint

Checklist

• I have read the CONTRIBUTING document.
• I have linked a related issue, or explained the problem above.
• I have added tests that prove my feature works.
• Ran gen-changesets skill, or this PR needs no changeset.
• Ran gen-docs skill, or this PR needs no doc update.

[changeset-bot]

🦋 Changeset detected

Latest commit: 1266d3c

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
@moonshot-ai/kimi-code	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: f390596f1f

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Preserve the effort override through OAuth login

When a fresh interactive startup reaches auth.login_required, this initial creation is abandoned and AuthFlowController.refreshConfigAfterLogin() later creates the session using thinkingEffortFromConfig(config.thinking) rather than startup.effort. Consequently, kimi --effort low silently loses the requested override whenever the user must log in first; forward the startup effort into the post-login creation path as is already done for the startup model.

Useful? React with 👍 / 👎.

[chatgpt-codex-connector]

Strictly validate effort on fresh sessions

For a fresh session, forwarding the value through createSession does not provide the model-specific validation promised by this flag: the v1 create path calls resolveThinkingEffort without strict Kimi-protocol validation, while the fresh v2 binding likewise omits strictThinking, so an unsupported value can be retained for a later upstream failure or silently normalized instead of rejecting the invocation. Resumed sessions do call setThinking and reject the same input, making behavior depend on whether the session already exists; fresh prompt, TUI, and native-v2 creation should use the strict setter/binding semantics too.

Useful? React with 👍 / 👎.