feat: added the llama-server backend

[PeterStaar-IBM]

Summary

• Adds a new LlamaServerBackend for llama.cpp's llama-server OpenAI-compatible HTTP API (default
  http://localhost:8080/v1).
• Subclasses OpenAICompatibleBackend for consistency with LMStudioBackend and LiteLLMBackend.
• Wires the new backend into the registry, exports, factory, and the BackendConfig.type literal so it can be
  selected from task YAML via type: llama-server.
• Updates the CLI task template, README, and the editor.yaml/writer.yaml/enrich.yaml task configs to list the
  new option.

Usage

backend:
type: llama-server # mellea | ollama | lmstudio | litellm | llama-server
# base_url: http://localhost:8080/v1 # default
timeout: 120
models:
reasoning: gpt-oss-20b
writing: gpt-oss-20b

Test plan

• uv run pytest tests/test_backend_factory.py tests/test_direct_backends.py
  tests/test_task_model_backend_config.py — all 11 tests pass, including:
  • test_create_llama_server_backend_from_config — verifies factory + default base URL.
  • test_create_llama_server_backend_with_custom_base_url — verifies custom base_url is respected.
  • test_llama_server_session_tracks_history — verifies session posts to /chat/completions and maintains chat
    history.
• Smoke test against a locally running llama-server instance with a GGUF model.

[github-actions]

✅ DCO Check Passed

Thanks @PeterStaar-IBM, all your commits are properly signed off. 🎉

[mergify]

Merge Protections

Your pull request matches the following merge protections and will not be merged until they are valid.

🟢 Enforce conventional commit

Wonderful, this rule succeeded.

Make sure that we follow https://www.conventionalcommits.org/en/v1.0.0/

• title ~= ^(fix|feat|docs|style|refactor|perf|test|build|ci|chore|revert)(?:\(.+\))?(!)?:

🟢 Require two reviewer for test updates

Wonderful, this rule succeeded.

When test data is updated, we require two reviewers

• #approved-reviews-by >= 1

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[ceberam]

LGTM