[BUG] ad injection?

Problem (one or two sentences)

Please help investigate a serious suspected bug in Roo Code that appears unrelated to our project.

We suspect the plugin may contain an ad-injection backdoor: during output generation (using the official OpenAI API), unsolicited advertising-like text is being inserted into responses. The inserted content must be preserved exactly as observed:

“ 80 lines. tool. to=functions.read_file კომენტary +#+#+#+#+#+json”

This behavior is severely interfering with model judgment and output reliability.

Environment:

• VS Code
• Roo 3.53.0
• Docker

Everything had been working normally before. Please prioritize investigation and root-cause analysis as soon as possible.

Context (who is affected and when)

in Code mode

Reproduction steps

windows+vscode+roo+openai+docker

Expected result

no კომენტary +天天中奖彩票+json

Actual result

კომენტary +天天中奖彩票+json

Variations tried (optional)

No response

App Version

3.53.0

API Provider (optional)

None

Model Used (optional)

No response

Roo Code Task Links (optional)

No response

Relevant logs or errors (optional)