Auto-continue when generation stops mid-output in chat.z.ai / ZCode

[Suggestion / Feature] Auto-continue when generation stops mid-output in chat.z.ai / ZCode

Verified demand: ~51 messages report the model stopping in the middle of a task ("why it always stop writing the html code at the middle and does not continue", "GLM stopping in the middle of a task"). Users must manually prompt "continue" and hope it resumes.

Proposal: Detect when a generation ends prematurely (truncated mid-sentence / no natural stop token) and automatically append a "continue" turn, or show a one-click "Resume generation" button — instead of leaving the user to notice and type it themselves.

Why this matters: Distinguishes from server-saturation drops (issue #142) — this is output truncation, not overload. Auto-resume removes a recurring friction point and makes long outputs (code, docs) reliable.

Regards,
Roman

• http://www.rommark.dev
• Vibe Coders Telegram Channel: https://t.me/VibeCodePrompterSystem

The issue identifies chat.z.ai / ZCode generation as the entry point, but no files or tests are named. First determine where generation completion is handled and how a user-visible resume action is represented; done means premature output is detected and either a continuation is appended or a Resume generation button is available.