LLM output is not being displayed when using Goose version 1.1.* with GitHub Models specified configuration. The problem persists despite validating the configuration file and doing a complete fresh reinstallation.

This configuration continues to work properly for Goose version 1.0.35

To Reproduce

Steps to reproduce the behavior:

1. Ensure the following configuration is used:

   OPENAI_HOST: https://api.githubcopilot.com
   OPENAI_CUSTOM_HEADERS: Copilot-Integration-Id=playground-dev
   OPENAI_BASE_PATH: chat/completions
   GOOSE_MODEL: gpt-4o
   extensions:
     computercontroller:
       bundled: true
       display_name: Computer Controller
       enabled: false
       name: computercontroller
       timeout: 300
       type: builtin
     developer:
       bundled: true
       display_name: Developer
       enabled: true
       name: developer
       timeout: 300
       type: builtin
     jetbrains:
       bundled: true
       display_name: Jetbrains
       enabled: false
       name: jetbrains
       timeout: 300
       type: builtin
     memory:
       bundled: true
       display_name: Memory
       enabled: false
       name: memory
       timeout: 300
       type: builtin
     tutorial:
       bundled: true
       display_name: Tutorial
       enabled: false
       name: tutorial
       timeout: 300
       type: builtin
   GOOSE_MODE: chat
   GOOSE_PROVIDER: openai

2. Open Goose.
3. Attempt any task that should display LLM output.
4. Observe the absence of output in the interface.

Expected behavior

LLM output should be displayed clearly when using Goose version 1.1.* with the given configuration.

Screenshots

Screenshot of Goose 1.1.3 behavior:

See that Goose is properly summarizing the input, however it is not actually displaying or handling any output.

Please provide the following information:

• OS & Arch: macOS Sequoia 15.5 M1 Max [e.g. macOS Ventura 13.0 ARM64]
• Interface: UI and CLI
• Version: 1.1.3
• Extensions enabled: None
• Provider & Model: OpenAI - gpt-4o (with GitHub Models)

Additional context

Add any other context about the problem here.