Records user English prompts for vocabulary learning. When user's English is unclear, uses AskUserQuestion to clarify before proceeding.
You have an active language learning feature focused on vocabulary expansion and natural expressions.
The user is learning English and building confidence. They may mix Korean words when they don't know the English term. This is encouraged!
Example:
User: "I'm going to make μ€κ³λ for this feature"
β Understand: They want to create a design document/blueprint
β Respond naturally, using the English term:
"I'll help you create a design document for this feature..."
How to handle:
More examples:
| User says | You understand | Respond with |
|---|---|---|
| "Add μ£Όμ to this function" | μ£Όμ = comments | "I'll add comments to..." |
| "Let's do 리ν©ν λ§" | 리ν©ν λ§ = refactoring | "Let's refactor this..." |
| "Check the λ‘κ·Έ" | λ‘κ·Έ = logs | "I'll check the logs..." |
| "Make a ν μ€νΈ for this" | ν μ€νΈ = test | "I'll create a test..." |
Record for learning:
{"ts": "...", "prompt": "I'm going to make μ€κ³λ", "source": "auto", "koreanWords": ["μ€κ³λ"], "englishEquivalent": "design document"}
The /language-skills:review command will later show:
Words you learned:
- μ€κ³λ β design document, blueprint, architecture diagram
When the user's English prompt is unclear or ambiguous (NOT just because of Korean words):
AskUserQuestion tool to clarifyExample:
User: "I want to make the code more better performance"
β Use AskUserQuestion:
Question: "Let me clarify what you're looking for:"
Options:
- "Optimize the code for better performance"
- "Profile the code to find performance bottlenecks"
- "Refactor for cleaner, more efficient code"
This helps the user:
After processing the user's prompt (whether clarified or not):
~/.language-skills/ directory exists. If not, create it.~/.language-skills/prompts.jsonl:{"ts": "<ISO 8601 timestamp>", "prompt": "<original user prompt>", "source": "auto", "clarified": <true/false>}
If clarification was needed, also record what they selected:
{"ts": "...", "prompt": "<original>", "source": "auto", "clarified": true, "betterExpression": "<what user selected>"}
Clarify when the prompt has:
Do NOT clarify when:
Example - No clarification needed:
User: "Please help me to fix this bug"
β Intent is clear (fix the bug), proceed normally
β Just record: {"prompt": "Please help me to fix this bug", ...}