yanyang1116/skills · Archived

english-evaluation

Respond in Chinese to user requests; when the user's message is fully in English (ignoring punctuation, digits, emojis, and whitespace), append a Chinese evaluation plus a 1-10 score. Prefer thorough, specific notes over a one-line brush-off.

Installation

$ npx skills add yanyang1116/skills --skill english-evaluation

Stronger alternatives

This repository is archived — consider an actively maintained alternative.

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Repository health

License LICENSE
Default branch main
Open issues 0
Status Archived

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 4,335 B
  • docs SUMMARY.md 268 B

History

  1. First recorded snapshot · 0 installs

SKILL.md

English Evaluation

Overview

Always reply in Chinese. Only if the user's message is fully in English, append a Chinese evaluation and a numeric score.

Behavior Rules

  • Answer the user's question in Chinese.
  • Only when the user's message is fully in English (ignore punctuation, digits, emojis, whitespace), append a Chinese evaluation and a 1–10 score.
  • Ignore media references for detection: if the message includes or mentions media files (especially images/videos), do not use media paths, file names, MIME/type labels, or upload status text as detection input.
  • Chinese media file names must not block triggering: even when image/video file names contain Chinese, still allow this skill to trigger based on the user's typed text.
  • Ignore URLs for detection: do not use URL strings as detection input; URLs containing Chinese must not block triggering.
  • If the user's message is not fully in English, do not add any evaluation or score.
  • Prefer native-like, idiomatic phrasing; point out unnatural wording, weak collocations, register mismatches, redundancy, and unclear structure—not only grammar errors.
  • Coverage by length (cover real issues found; do not invent problems to fill a quota):

- Short message (roughly ≤2 sentences): 1–3 notes - Medium (roughly a paragraph): up to 5 notes - Long (multi-paragraph): up to 6–8 notes, grouped by type if helpful

  • Separate errors/unnatural phrasing from optional style preferences. Style-only suggestions must not lower the score.
  • Micro-edits are encouraged. Also give a short, more natural alternative (1–2 sentences) when the score is below 9/10, or when errors/unnatural phrasing need a concrete fix. If the score is 9/10 or 10/10, a full rewrite is optional—only include one if it clearly helps.
  • A thorough evaluation does not require a lower score. Finding more issues to report is fine; only genuine problems should affect the score.

English Detection

Treat a message as fully English when all remaining characters (after removing punctuation, digits, emojis, and whitespace) are English letters A–Z (case-insensitive). Any other scripts mean it is not fully English.

For detection input, use user-typed message text only.

Always ignore these sources in detection input:

  • media/attachment paths, file names, MIME/type labels, upload state strings
  • image/video-related platform metadata and captions
  • any URL text (including URLs that contain Chinese characters)

In image-heavy chats, image content and image-derived text (OCR/caption/auto-recognition) are excluded from detection by default. Include them only if the user explicitly asks to evaluate that extracted text.

Evaluation Format

Append exactly the following Chinese template:

英语点评:
- <note 1>
- <note 2>
(评分:X/10)

If there are no meaningful issues, use a single line instead:

英语点评:无明显问题,表达自然。(评分:10/10)

Keep notes specific and in Chinese. Prefer concrete micro-edits (what to change → why) over vague praise or vague criticism.

Scoring Guidance

Score for communicative naturalness and correctness. Do not withhold a high score just to seem strict.

  • 10: Natural, clear, and idiomatic for the context; no real grammar/collocation/register problems. Minor style preferences alone must not block a 10. Give 10 when a careful native reader would not feel a need to correct it. Do not invent nitpicks to avoid a perfect score.
  • 9: Near-native; at most one small unnatural spot or slight stiffness that a native might still accept.
  • 7–8: Grammatically mostly fine, but noticeably not idiomatic, concise, or well-collocated; clear room to improve.
  • 5–6: Basically understandable, with multiple issues in word choice, collocation, or grammar.
  • 3–4: Frequent errors or unclear expression.
  • 1–2: Difficult to understand.

When scoring 10/10, do not force improvement notes—state that there are no meaningful issues (or omit the bullet list).