valid-unicode_skill

Overview

This skill validates that agent inputs, outputs, and prompts correctly handle Unicode characters and emoji. It provides targeted checks and examples to ensure text encoding, rendering, and round-trip integrity remain intact across pipelines. Use it to catch encoding regressions and to certify multilingual and emoji-rich content flows.

How this skill works

The skill runs a suite of lightweight checks against sample strings containing Unicode characters, accented letters, CJK text, and a variety of emoji. It inspects encoding handling, preserves grapheme clusters, verifies serialization/deserialization round trips, and flags replacement characters or truncation. Results summarize failures and point to the offending stage (input, model I/O, or storage).

When to use it

• When onboarding new LLMs, adapters, or runtime environments to confirm Unicode fidelity
• When adding multilingual prompts or user content that includes non-Latin scripts
• When supporting emoji reactions, rich text, or social-media style content
• When debugging lost characters, � replacement marks, or incorrect text slicing
• When validating data pipelines that store or transmit user text

Best practices

• Include representative samples from all target languages and common emoji sequences
• Test both single-character and compound grapheme clusters (emoji modifiers, flags)
• Verify both UTF-8 encoding and storage backends (databases, message queues, file systems)
• Run checks in CI to catch regressions early and after deploys
• Log raw byte sequences when diagnosing mysterious character corruption

Example use cases

• Preflight tests before switching LLM providers to ensure prompt fidelity with CJK, Arabic, and emoji
• CI check that detects when a serialization change introduces replacement characters
• Debugging a chat system where emoji modifiers or skin tones are lost or split incorrectly
• Validation for analytics pipelines that aggregate user text from multiple locales