ci: run Python runtime tests in CI (#945)

[himanshu231204]

Summary

Adds a dedicated uv-based Python CI job that runs the prime-agent-runtime test suite across all supported Python versions, and connects it to the aggregate CI success gate so Python harness/runtime failures block merges and releases.

Fixes #945

Problem and impact

Python harness/runtime regressions can merge and ship without any direct Python test executing. The CI graph (.github/workflows/ci.yml) contained only Node workspace/package lanes; prime-agent-runtime/test/test_*.py was never invoked by any job. A failure confined to a Python test expectation would go completely unobserved.

Changes

.github/workflows/ci.yml

• Added a new python-runtime job:
  • Matrix over supported Python versions 3.10, 3.11, 3.12, 3.13 (matches requires-python = ">=3.10" in pyproject.toml).
  • Uses astral-sh/setup-uv@v9.0.0 with python-version and enable-cache for fast, cached uv installs.
  • Runs uv sync --frozen (deterministic install from the committed lockfile) then uv run pytest.
  • Runs from the repo root checkout so the cross-language artifact contract test (test_agent_message_skill.py, which imports the shipped skill from packages/coding-agent/skills/agent-message/) resolves correctly.
• Wired python-runtime into the build-check-test aggregate gate: added it to needs and to the result check (PYTHON_RUNTIME_RESULT). Failures now block aggregate CI and release, rather than being hidden in an unrelated kernel smoke test.

prime-agent-runtime/pyproject.toml

• Added [dependency-groups] dev group: httpx, mcp>=1.0, pytest>=8.0.
  • httpx is required because src/rlm/mcp_base.py imports it for the http_client transport signature (mcp 2.x style). Production always installs httpx as a default kernel extra package (DEFAULT_RLM_EXTRA_PACKAGES in packages/coding-agent/src/core/kernel/bootstrap.ts), so the test environment now mirrors the production package set.
  • mcp is required by test_mcp_base.py (it patches mcp.ClientSession and the streamable-HTTP transport).
• Added [tool.pytest.ini_options] with testpaths = ["test"].

prime-agent-runtime/uv.lock

• Committed the generated lockfile so uv sync --frozen installs dependencies deterministically in CI and locally. Dependabot already tracks /prime-agent-runtime for the uv ecosystem.

.gitignore

• Added .venv/ and .pytest_cache/ (created by uv sync / pytest runs).

Verification

• uv sync --frozen && uv run pytest → 64 passed on Python 3.11.
• Same suite → 64 passed on Python 3.13.
• uv lock --check → lockfile in sync.
• YAML lint → valid.
• Pre-commit npm run check → passed.

Acceptance criteria coverage

• Supported Python versions run the runtime suite in CI — matrix 3.10–3.13.
• Failures block aggregate CI and release — python-runtime is a dependency of the build-check-test aggregate gate.
• Dependencies installed deterministically — committed uv.lock + uv sync --frozen.
• Cross-language artifact contract covered — test_agent_message_skill.py imports the shipped skill from the TypeScript package's skills/ directory.

Notes

• The pkg:coding-agent label is not yet provisioned on the repo (per the issue audit metadata); bug and github_actions labels should be applied by a maintainer.

[Copilot]

Pull request overview

Adds Python runtime test coverage to the repo’s CI by introducing a dedicated uv/pytest job for prime-agent-runtime and wiring it into the aggregate CI gate so Python regressions block merges/releases.

Changes:

• Added a python-runtime GitHub Actions job that runs the runtime test suite on a Python 3.10–3.13 matrix using uv.
• Added a dev dependency group plus pytest discovery configuration to prime-agent-runtime/pyproject.toml.
• Ignored local Python virtualenv and pytest cache directories in .gitignore.

Reviewed changes

Copilot reviewed 2 out of 4 changed files in this pull request and generated no comments.

File	Description
.github/workflows/ci.yml	Adds a python-runtime matrix job and gates build-check-test on its result.
prime-agent-runtime/pyproject.toml	Adds dependency-groups.dev and configures pytest to look under test/.
.gitignore	Ignores .venv/ and .pytest_cache/ to reduce local noise.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[sethkarten]

Thank you for the report and proposed work. This root cause is now covered by maintainer-owned stacked PR #1158, authored independently from upstream/main.

We did not inspect or reuse this PR's diff, branch, commits, implementation code, or tests; its public description/comments were used only as a bug report. To keep one review surface, this PR is superseded by #1158 and is being closed.

The complete review stack is #1158–#1165. It is being left unmerged for human review after CI and review-bot findings are cleared.