CodeContinue v2.5.2
===================

🚀 FIRST THINGS FIRST: RUN THE SETUP WIZARD
--------------------------------------------
1. Press: Ctrl+Shift+P  (or Cmd+Shift+P on macOS)
2. Type:  CodeContinue: Configure
3. Press: Enter

The step-by-step setup wizard will configure everything in under 30 seconds:
  1. Completion Endpoint & Model (auto-detects models from server)
  2. Dual-Engine Chat Endpoint & Model (local runner or cloud like Claude/GPT)
  3. 1-Click Keybinding Setup (Recommended: Ctrl+Enter Suggest, Tab Accept, Esc Dismiss)


Patch Release - Markdown Commentary & Fence Stripping:

1. Clean Code Extraction:
   • Fixed an issue where models (e.g. Qwen 2.5 Coder, Claude, GPT-4o) appending
     conversational explanations (e.g. "This code snippet demonstrates...") after
     closing markdown fences would leak the backticks and commentary into the editor.
   • Upgraded clean_markdown_fences to cleanly isolate code blocks and strip all
     leading greetings, trailing commentary, and markdown code fences.

2. Negative Constraint Reinforcement:
   • Reinforced prompt templates in Suggest and Edit to explicitly instruct
     models to output raw code without backticks or conversational prose.

Questions or issues?
https://github.com/kXborg/codeContinue/issues
