test(e2e): add end-to-end tests for inference and CLI

[doringeman]

Adds an end-to-end test suite that builds the full stack from source and runs inference and CLI tests against a live model-runner instance. The test harness follows the same pattern as the dmr convenience wrapper: find a free port, start the server pointing at the local llama.cpp build, wait for health, run tests, tear down.

make e2e

In general, the fix is to explicitly declare a permissions block that grants only the minimal required scopes to the GITHUB_TOKEN. Since this job only checks out repository contents and runs tests, it only needs read access to repository contents.

The best fix here is to add a permissions block to the e2e-test job (or at the workflow root). To keep the change tightly scoped and avoid affecting other workflows, we will add it at the job level, immediately under e2e-test: and aligned with runs-on:. We will set contents: read, which is sufficient for actions/checkout and normal test execution and preserves existing functionality while constraining token capabilities.

Concretely, in .github/workflows/e2e-test.yml, modify the e2e-test job definition so that:

• After line 11 (e2e-test:), insert:

    permissions:
      contents: read

• Keep all existing steps and configuration unchanged.

No additional methods, imports, or definitions are needed.

[sourcery-ai]

Hey - I've left some high level feedback:

• The GitHub workflow is configuring go-version: 1.25.8, which is not a valid Go release; this should be updated to a real supported version (e.g., 1.22.x or 1.23.x) to avoid CI setup failures.
• make e2e already depends on build-llamacpp and build, but TestMain also runs make build, which is redundant and will slow down CI; consider relying on the pre-built binaries in CI or removing the extra build step from TestMain.

Prompt for AI Agents

Please address the comments from this code review:

## Overall Comments
- The GitHub workflow is configuring `go-version: 1.25.8`, which is not a valid Go release; this should be updated to a real supported version (e.g., 1.22.x or 1.23.x) to avoid CI setup failures.
- `make e2e` already depends on `build-llamacpp` and `build`, but `TestMain` also runs `make build`, which is redundant and will slow down CI; consider relying on the pre-built binaries in CI or removing the extra build step from `TestMain`.

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}

[gemini-code-assist]

Code Review

This pull request introduces a comprehensive suite of end-to-end (e2e) tests for the model-runner. It adds new Makefile targets to facilitate building the llama.cpp submodule and executing these e2e tests. The tests cover both CLI functionality (pull, list, run, remove) and API-based inference (pull, list, chat, streaming, remove). A review comment highlighted an issue in the inference tests where io.ReadAll errors are ignored, which could mask underlying I/O problems and lead to confusing test failures.