Claude Sonnet 5 Bug Causes 100% Hangul Corruption in Tool Calls
whchoi98 · ghdev · 2026-08-01
A developer has root-caused a frequently reported Korean (Hangul) corruption issue in Claude Sonnet 5. Through a controlled A/B experiment, they discovered that the model sometimes writes Korean tool-parameter strings as \uXXXX unicode escape sequences instead of literal UTF-8, while simultaneously mis-spelling the hex codepoints.
This results in valid-but-wrong Hangul syllables, producing nonsensical text. The experiment revealed a 100% corruption rate (45/45 runs) whenever the model engaged in escape-writing, whereas literal UTF-8 runs were mostly clean. The developer noted that a simple output-format instruction in the system prompt can fully suppress the issue.
More from coding & agent
- DeepMind launches Antigravity for Science: coding agent built for end-to-end research workflows — pushmeet · 2026-10-16
- DeepMind turns coding agent Antigravity into a science tool, surfacing novel antimicrobial peptide in one session — pushmeet · 2026-10-16
- Cross-review coding workflow: Claude writes the plan, Codex audits it — theteknosaur · 2026-10-16
- Simon Willison: code mode is the future of tool calling, if it can be sandboxed — springrod · 2026-10-16
- Normalizing Claude's schema-free MCP output before it hits downstream n8n nodes — snowingbol · 2026-10-16
- Agent named 'Ilya' reportedly scores 1.1pt higher on benchmarks — daniel_mac8 · 2026-10-16