Preface

In LLM applications, “English writing correction” is often underestimated. Simply asking the model to “help me correct my essay” often results in generic polishing: grammar errors aren’t fully flagged, Chinglish isn’t pointed out, and grading criteria remain unclear. To implement this in a real product, you need to decouple syntax annotation, expression enhancement, logical coherence, and scoring rubrics into distinct steps, while also covering the specific requirements of CET-4, CET-6, and Postgraduate Entrance Exam English (I/II).

The “English Essay Correction” Skill introduced below is a packaged solution in this direction: released by the China Mobile Hangzhou R&D Center on SkillHub, it converges the aforementioned correction workflow into a callable capability.

What is this

One-sentence positioning: A correction Skill designed for writing and translation tasks in domestic English exams. It takes essay or translation text as input and outputs sentence-by-sentence corrections, scores, optimized drafts, and score-improvement suggestions.

  • Name: English Essay Correction
  • Identifier: @org-36nomya8/sk-bd1431891fc0
  • Version: 1.0.0
  • Source: enterprise
  • Category: education
  • Tags: requires_api_key=false (no additional API Key configuration required)
  • Publisher: China Mobile Hangzhou R&D Center

The problem it solves: When an upper-layer Agent or application receives English essays, chart-based essays, letters/notices, postgraduate exam essays (Type A/B), CET-4/6 writing, or translation texts pasted by users, it enables the system to provide interpretable correction results based on exam rubrics, rather than a paragraph of generic revision suggestions.

Core Functions

According to the introduction on the SkillHub page, this Skill includes the following capabilities:

  1. Performs sentence-by-sentence correction for CET-4, CET-6, and Postgraduate Entrance Exam English (I/II) essays or translations, rather than providing a single summary opinion for the whole text.
  2. Flags five categories of issues: grammar, spelling, punctuation, vocabulary collocation, and Chinglish. Specifically, “Chinglish”—the most common cause of point loss in domestic exam writing—is listed separately.
  3. Enhances appropriate academic expressions, supplements logical transitions, and breaks up lengthy sentences (text rewriting and polishing).
  4. Identifies writing strengths and scores them according to the specific exam’s rubric; grading is based on standardized criteria, not arbitrary judgment.
  5. Provides optimized drafts on the same topic, verifiable excellent model essay information, and score-improvement suggestions as follow-up actions after correction.

Note: The capabilities listed above are official descriptions. Actual effectiveness depends on the underlying model and specific input; it is recommended to verify with sample texts yourself.

Installation and Activation

This Skill is based on the general SKILL.md format and has no additional dependencies. It is tagged with requires_api_key=false, so no separate key configuration is needed.

  • Official API URL: https://api.skillhub.cn/org-36nomya8/sk-bd1431891fc0
  • SkillHub Page: https://www.skillhub.cn/skills/org-36nomya8/sk-bd1431891fc0

Regarding installation commands for Cursor, Codex CLI, and Claude Code, the retrieved materials do not provide them, so no speculation is made here. If needed, please open the SkillHub page above and follow the platform-specific guidance, or consult SkillHub’s general installation documentation.

Typical Usage

The materials do not provide official example invocation code. The usage methods below are organized based on the capabilities described in the introduction; actual invocation depends on your Agent environment.

User Input Form: Pasted or uploaded text, including:
1. English essays (including CET-4/6 writing, Postgraduate exam essays Type A/B)
2. Chart-based essays
3. Applied writing (letters, notices, etc.)
4. Translation text

Trigger Scenario: When the user explicitly requests one of the following actions: correction, polishing, scoring, rewriting, model essays, logic optimization, or expression enhancement.

Invocation Recommendation: It is recommended to pass the exam type (CET-4 / CET-6 / Postgraduate English I / Postgraduate English II) as context so that the Skill can grade according to the corresponding rubric.

Applicable Scenarios and Caveats

Suitable Scenarios: Writing practice products for domestic English exams, homework correction assistance, translation assignment grading, and educational Agents requiring scores and improvement suggestions based on exam rubrics.

Key Caveats:
1. Scope Limitation: Applicable only to CET-4, CET-6, and Postgraduate Entrance Exam English (I/II). IELTS, TOEFL, GRE, etc., are not declared in the introduction; do not apply this Skill’s rubrics to them directly.
2. Focus: It handles writing and translation responses, not general English Q&A, speaking, or listening.
3. Score Reliability: Scoring results depend on how well the underlying model executes the exam rubric. Since this involves subjective grading, treat the results as “reference feedback” rather than the sole basis for judgment.
4. Version Status: Skill version is 1.0.0, source is enterprise, publisher is China Mobile Hangzhou R&D Center. The SkillHub page retrieval showed a rating of 100000, 1222 downloads, and 0 installs, suggesting there may be no large-scale usage data available for reference yet.
5. Configuration: Regarding installation commands and specific platform configuration, the retrieved materials do not cover this. Please refer to the official page before use.

Conclusion

The value of this Skill lies in transforming “English writing correction” from a one-off model prompt into a structured process featuring syntax annotation, expression enhancement, rubric-based scoring, and optimized drafts. It is suitable for direct integration into educational Agents. If you are dealing with CET-4/6 or Postgraduate English writing and translation, you can start experimenting here; if you are targeting other examination systems, you need to confirm the rubric compatibility first.

Official URL: https://api.skillhub.cn/org-36nomya8/sk-bd1431891fc0