[LivXue]

Summary

Synchronize the Chinese README with the updated English overview. The change aligns the Raven introduction and built-in agent section, including the Host Agent positioning, Raven Evolver description, built-in agent wording, and orchestration readiness statement.

The existing benchmark charts, links, captions, and the remaining bilingual sections are preserved.

Type

• Fix
• Feature
• Docs
• CI / tooling
• Refactor
• Other

Verification

• Relevant tests pass locally
• Relevant lint / type checks pass locally
• User-facing docs or screenshots are updated when needed

Commands run:

• uv run --no-sync pytest tests/test_readme_scope_canon.py tests/test_cli_onboard_commands.py -k readme_quickstart -q (1 passed)
• uv run --no-sync pytest tests/test_docker_runtime.py -k container_exposes_no_engine_selector -q (1 passed)
• UV_NO_SYNC=1 make check-large-files check-source-language PYTHON_VERSION=.venv/bin/python COMMIT_RANGE=HEAD (passed)
• git diff --check (passed)

Risk

• Security impact considered
• Backward compatibility considered
• Rollback path is clear for risky changes

This is a documentation-only change. Reverting the commit restores the previous README wording.

Related Issues

N/A

[gloryfromca]

Blocking: the overview contradicts the documented Evolver architecture.

I found one blocking documentation error, marked inline.

Coverage: I reviewed the full bilingual diff and surrounding README sections; checked AGENTS.md and CONTEXT-MAP.md; traced shipped-agent discovery and onboarding behavior; checked the Evolver boundary and import contract; examined recent README history and the PR record; considered link/heading compatibility; and confirmed that no tests were weakened.

Verification: the two targeted pytest selections each passed (1 passed); check_large_files.py, check_source_language.py, and git diff --check passed; the worktree is clean; and all GitHub CI checks are green. The first local --no-sync attempt found no prepared pytest environment, so I synchronized the required Python 3.12/all-extras environment and reran successfully.

[ZuyiZhou]

Approving: docs only, two files, all 18 checks green, and the bilingual halves stay in step.

Checked the zh text against the new en text line by line -- the Host Agent framing, the EverOS sentence, the built-in agent line, the Raven Evolver paragraph and the new blockquote all carry the same claims in both files, and the same two paragraphs are dropped on both sides. The benchmark image, its link, the surrounding sections and the repo-layout table are untouched, so tests/test_readme_scope_canon.py keeps its anchor and nothing keys on the changed H1.

Two notes, neither blocking, both for a follow-up if you agree:

1. The new blockquote "All four agents are built in and ready for orchestration out of the box" sits in tension with the install section this PR leaves alone, which still says "Setup asks about each product and registers the ones you take up" (README.md:62-66, README.zh-CN.md:62). That paragraph matches the code: agents carry a per-agent enabled flag and setup prompts for each one. The removed line "Enable the agents you need during onboarding" was the accurate half. The agents/README.md link survives at README.md:66, so nothing is orphaned, but the two statements now read differently on the same page.

2. "state-of-the-art (SOTA) performance in their respective domains" is a strong public claim across four domains, and the README shows one benchmark, for orchestration. Worth either scoping the claim to what the chart supports or pointing at per-agent numbers.

Style nit: the en README had no curly apostrophes and no em dashes before this change (8 straight 's, 0 of either); it now gains EverMind's with a curly apostrophe and one em dash. The zh -- is idiomatic there and fine. No gate rejects it, since *.md self-governs, but the en file is otherwise ASCII-punctuation throughout.