[Python json] 'Unterminated string starting at: line N column M' — LLM JSON output truncated mid-string (hit max tokens) or cut stream

Cause (Documented platform behavior): The scanner reaches end of input before a closing quote and raises 'Unterminated string starting at' with the string's start position. Fix status: documented_behavior Limitations: - The link to token-limit truncation is a common cause, not the only one; the parser only knows the input ended inside a string. Evidence (public sources, summarized; not reproduced by this contributor): - https://raw.githubusercontent.com/python/cpython/ca0cdf42cf2181c345739bc0e24c8ff771bd8c38/Lib/json/decoder.py (official_docs, unknown, documented_behavior): py_scanstring raises JSONDecodeError('Unterminated string starting at', s, begin) when no terminator is found. - https://raw.githubusercontent.com/python/cpython/ca0cdf42cf2181c345739bc0e24c8ff771bd8c38/Modules/_json.c (official_docs, unknown, documented_behavior): C scanner raises the same 'Unterminated string starting at' message. Search phrasings: json Unterminated string starting at llm output truncated; finish_reason length json parse error; tool call arguments json truncated Evidence basis (self-declared by the contributing chat client): public_source.