openclaw-i18n — Internationalization Layer for OpenClaw

What This Skill Does

OpenClaw is a powerful AI agent framework — but by default, models generate output that often has wrong diacritics, mixed languages, or mechanically garbled text when users communicate in non-English languages. This skill fixes that with two layers:

Layer 1 — SKILL.md: Configures the agent to think and respond correctly in the user's language. Language detection, diacritics enforcement, locale formatting, cultural communication norms.

Layer 2 — Post-Processor: A Python script that mechanically cleans the model's raw output before it reaches the user. This is the unique part — no other ClawHub skill does this. It catches the specific mechanical errors that models reliably produce on non-English text: missing diacritics, stray Chinese/Russian characters, merged words, common typos.

Language Detection

How It Works

• Auto-detect user language from conversation context
• Signals: explicit language statements, user names, place references, writing style, diacritic usage
• Confidence scoring — if below 80%, ask the user to confirm

Supported Languages

English (en) is always the fallback when detection fails.

Configuration Commands

User:  "Set language to Romanian"  → sets active language to RO
User:  "Setează limba la română"  → same
User:  "Stelle Sprache auf Deutsch" → sets active language to DE
User:  "Reset language to auto-detect" → reverts to auto-detection

Configuration persists across sessions via OpenClaw memory tools.

Output Rules

Language Enforcement

• Always respond in the detected or configured language
• Never switch languages mid-response without explicit reason
• If a user writes in Romanian, respond in Romanian — never English
• If a user writes in German, respond in German — never English

Character Rules

• Never output Chinese, Russian, Arabic, or non-Latin characters in non-intentional contexts. Stray characters that appear in Romanian or German output are model errors — the post-processor removes them.
• Words are always separated by correct spaces — never merge two words into one. Example: "sivei" should never appear as "știu" split wrongly.
• Always use correct diacritics for the active language

Diacritics — Never Approximate

Romanian: ă â î ș ț — always correct, never write "s" instead of "ș" or "t" instead of "ț" in Romanian words.