File edits include the model/Continue's reasoning/thinking

[evanshortiss]

Before submitting your bug report

• I've tried using the "Ask AI" feature on the Continue docs site to see if the docs have an answer
• I'm not able to find a related conversation on GitHub discussions that reports the same bug
• I'm not able to find an open issue that reports the same bug
• I've seen the troubleshooting guide on the Continue Docs

Relevant environment info

- OS: macOS
- Continue version: 1.2.16
- IDE version: Red Hat OpenShift Dev Spaces with Microsoft Visual Studio Code - Open Source IDE Version: 1.104.3
- Model: qwen3-14b
- config:
  
version: 0.0.1
models:
  - name: VLLM (qwen3-14b)
    provider: vllm
    model: qwen3-14b
    apiKey: $VLLM_API_KEY
    apiBase: $VLLM_HOST/v1
    roles:
      - chat
  
  OR link to agent in Continue hub:

Description

Since updating to Continue plugin 1.2.16 the model's thought process is injected into files it edits. This screenshot contains a sample. Downgrading to a prior version resolves the issue.

To reproduce

Use Continue 1.2.16 with a vLLM hosted model. I have not yet confirmed if this occurs with other providers.

Log output