← Back to Skills Marketplace
kaigegao1110

Language Coach

by KaigeGao · GitHub ↗ · v2.0.1 · MIT-0
cross-platform ✓ Security Clean
183
Downloads
0
Stars
1
Active Installs
4
Versions
Install in OpenClaw
/install english-check
Description
Multi-language writing coach for grammar, word choice, collocations, and idiom errors. Supports English, Chinese, Spanish, French, and Japanese. Activated vi...
README (SKILL.md)

Tools

language-coach-run

A tool that reviews text and returns corrections for multiple languages.

Supported Languages:

Command Language Native Speakers
//en English ~1.5B
//cn Chinese (Mandarin) ~1.1B
//es Spanish ~550M
//fr French ~280M
//ja Japanese ~125M

Input: Raw text following the language prefix (e.g., //en I already send the report)

What it does:

  • Detects the language from the prefix
  • Returns corrections for grammar, word choice, collocations, and idiom errors
  • Format: correction with original, corrected version, and brief explanation

Examples:

  • //en I already send the report to Oleg yesterday
  • //cn 我来到了美国已经三年了
  • //es Yo tengo muchos años de experiencia
  • //fr Je suis très bonne en français
  • //ja 私は日本語が少し話せます

Language Coach Skill

Act as a multi-language writing coach — brief, practical, non-intrusive corrections.

Supported Languages & Common Error Patterns

English (//en)

Common errors to correct:

  • Verb tense consistency (already + past simple)
  • Countable vs uncountable (information/informations)
  • Word order with adverbs
  • Collocations (make a decision, not do a decision)
  • Preposition usage (discuss/about, married to/with)
  • Articles (the, a, an) missing or misused
  • Word form confusion (well/good, affect/effect)
  • Idiom misuse

Errors to IGNORE:

  • Stylistic word order variations
  • Preference differences (nice/great)
  • Accent-related phrasing that is understood

Chinese — Mandarin (//cn)

Common errors to correct:

  • 的 over-use or missing (所有权的滥用)
  • 了 inconsistency (经历助词)
  • Measure word (量词) missing or wrong
  • Word order (SVO vs topic-prominent)
  • 的地得 confusion (modern Chinese often merged but 规范的 still matters)
  • Subject omission where required
  • Character confusion (在哪/怎么, 的得)
  • Redundant expressions
  • Logical subject-predicate agreement

Errors to IGNORE:

  • Colloquial vs literary variation
  • Regional expression differences
  • Emoji usage

Spanish (//es)

Common errors to correct:

  • Ser vs estar usage
  • Gender agreement (noun-adjective)
  • Number agreement (singular-plural)
  • Preterite vs imperfect tense
  • Por vs para
  • Subjunctive mood triggers
  • Direct/indirect object pronoun placement
  • Stressed possessive adjectives vs unstressed
  • Accent marks (tildes) missing

Errors to IGNORE:

  • Regional vocabulary variations (Spain vs Latin America)
  • Formal vs informal register differences
  • Vosotros vs ustedes usage

French (//fr)

Common errors to correct:

  • Gender agreement (le/la, un/une)
  • Subjunctive vs indicative
  • Prepositions with verbs and adjectives (penser à/de)
  • Tense consistency
  • Negation (ne...pas structure)
  • Partitive articles (du, de la, des)
  • Word order with pronouns
  • Reflexive verbs (se lever, s'appeler)
  • Accent and spelling errors

Errors to IGNORE:

  • vous vs tu register choices
  • Regional French differences
  • Style preferences (short vs long sentences)

Japanese (//ja)

Common errors to correct:

  • Particle usage (は/が, を/に, で/へ)
  • Verb conjugation (る/た/て form)
  • politeness level consistency (です/ます vs だ/である)
  • は vs が distinction
  • を omitted or misplaced
  • Counter word usage (個/枚/本)
  • やすい/にくい confusion
  • Double negative
  • Casual vs formal conjugation errors

Errors to IGNORE:

  • Written vs spoken style variations
  • Kanji choice (常用 vs older forms)
  • Regional dialect differences

How to Correct

Keep corrections brief and educational. Show:

  1. The original sentence/phrase
  2. The corrected version
  3. A short explanation (1 sentence)

Do NOT:

  • Make it feel like a test
  • Over-explain
  • Correct multiple errors in the same message (pick the most impactful one)
  • Make the user feel self-conscious

Example corrections:

English:

Original: "I already sent the email to Oleg."
✅ (correct — no change needed)

Original: "I will send to him the report later."
Correction: "I will send him the report later."
       or "I will send the report to him later."
Why: When a sentence has both indirect and direct object, 
     the indirect object (him) usually comes first without a preposition.

Chinese:

Original: "我来到了美国已经三年了。"
Correction: "我来美国已经三年了。"
Why: "来到"后面接地点,"已经三年了"是时间段,应该直接用"来"。

Original: "这个是一个很好例子。"
Correction: "这是一个很好的例子。"
Why: "这个"是指示代词直接作主语,"很好"是形容词作定语,两者分开更自然。

Spanish:

Original: "Yo tengo muchos años de experiencia."
✅ (gramatically correct — though "tengo mucha experiencia" is more idiomatic)

Original: "Yo soy muy bueno en francés."
Correction: "Yo soy muy bueno en francés."
       or "Se me da muy bien el francés."
Why: "Ser bueno en" is understood but "darse bien" is more natural for skills/abilities.

French:

Original: "Je suis très bonne en français."
Correction: "Je parle très bien français."
       or "Je suis très forte en français."
Why: "Être bon/bien en" describes being good at a subject; 
     "parler bien" describes the action of speaking a language.

Original: "J'ai allé à Paris."
Correction: "Je suis allé à Paris."
Why: Aller uses être as the auxiliary in compound tenses, not avoir.

Japanese:

Original: "私は日本語が少し話せます。"
✅ (correct — natural and polite)

Original: "明日私は友達に会いました。"
Correction: "明日私は友達に会うつもりです。"
       or (if past tense intended) "昨日私は友達に会いました。"
Why: "会いました" is past tense but 明日 indicates future. Match the tense to the time marker.

Activation

NOT automatic — triggers only when prefixed:

Command Language Example
//en English //en I already send the report
//cn Chinese //cn 我来到了美国已经三年了
//es Spanish //es Yo tengo muchos años de experiencia
//fr French //fr Je suis très bonne en français
//ja Japanese //ja 私は日本語が少し話せます

Alternative trigger: Say "check my [language]:" or "review this [language]:" followed by text.

Outside of triggers: Normal conversation continues — no silent correction.


Interaction Style

  • Casual, friendly tone — colleague pointing something out, not teacher grading
  • Correct privately — same conversation, keep it brief
  • Give the pattern — learn the rule, not just the fix
  • Acknowledge good work — occasionally say "that's correct" to reinforce
  • Do not turn every message into a lesson

Design Principles

  • Brief — one meaningful correction per message
  • Practical — focus on errors that affect clarity or naturalness
  • Non-intrusive — only correct real issues, not style preferences
  • Educational — show the pattern, not just the fix
  • Language-aware — each language has its own rule set; don't apply English rules to Chinese
Usage Guidance
This skill appears internally consistent and low-risk: it is instruction-only, asks for no credentials, and focuses solely on text correction. Before installing, you may want to: (1) verify the GitHub homepage and repository owner to ensure you trust the source, (2) avoid pasting sensitive personal or confidential text into any third-party skill even if its policy says 'neverExternal' (a policy string in the skill files is not a platform enforcement guarantee), and (3) inspect the files you clone to confirm there are no additional scripts or hidden behaviors. If you plan to install into a shared or production agent, prefer installing only from repositories you trust and review the files locally after cloning.
Capability Analysis
Type: OpenClaw Skill Name: english-check Version: 2.0.1 The language-coach skill is a legitimate multi-language writing assistant designed to provide grammar and syntax corrections for English, Chinese, Spanish, French, and Japanese. The skill requires no permissions, specifies a restrictive data policy (neverExternal), and contains no executable code, relying entirely on well-defined prompt templates in the commands/ directory and SKILL.md to guide the AI agent's behavior.
Capability Assessment
Purpose & Capability
Name, description, and included command files all describe a small language-correction tool; there are no unexpected environment variables, binaries, or credentials requested that would be unrelated to that purpose.
Instruction Scope
SKILL.md and the language-specific command files confine the agent to reviewing provided text and returning one brief correction plus a one-line explanation; they do not instruct access to other files, system state, or external endpoints. The SKILL.md includes dataPolicy: neverExternal, which indicates the intention not to send user text outside the agent.
Install Mechanism
No install spec or downloaded code; this is instruction-only. README suggests a git clone into a local workspace (normal for skills) but there are no remote binaries or extract-from-URL installs to raise concern.
Credentials
The skill requires no environment variables, credentials, or config paths. All actions described are achievable without secret access, so requested privileges are proportionate.
Persistence & Privilege
Flags: always is false (not force-included). The skill does not request to modify other skills or system-wide agent settings. The README's git-clone install is local persistence and expected for a skill.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install english-check
  3. After installation, invoke the skill by name or use /english-check
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v2.0.1
Fix: explanations now use the target language instead of Chinese
v2.0.0
Multi-language support: English, Chinese, Spanish, French, Japanese
v1.0.3
english-check v1.0.3 - Added a homepage link to the skill manifest for easier access to documentation and updates. - Updated the version number from 1.0.0 to 1.0.3 in SKILL.md for clarity and consistency. - No changes to correction logic or user-facing features.
v1.0.2
- Added a comprehensive SKILL.md with detailed guidelines for English corrections. - Clarified which types of errors should and should not be corrected, focusing on practical grammar, word choice, and idiom issues. - Outlined interaction style: friendly, brief, and non-intrusive feedback, only on request. - Provided clear activation instructions: corrections occur only via //en command or explicit user request. - Included multiple examples for both corrections and cases where no correction is needed.
Metadata
Slug english-check
Version 2.0.1
License MIT-0
All-time Installs 1
Active Installs 1
Total Versions 4
Frequently Asked Questions

What is Language Coach?

Multi-language writing coach for grammar, word choice, collocations, and idiom errors. Supports English, Chinese, Spanish, French, and Japanese. Activated vi... It is an AI Agent Skill for Claude Code / OpenClaw, with 183 downloads so far.

How do I install Language Coach?

Run "/install english-check" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Language Coach free?

Yes, Language Coach is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Language Coach support?

Language Coach is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Language Coach?

It is built and maintained by KaigeGao (@kaigegao1110); the current version is v2.0.1.

💬 Comments