ci: add Codecov coverage reporting

[cmeans-claude-dev]

Summary

Mirrors mcp-clipboard's Codecov integration. Requested by @cmeans after the mcp-* alignment planning — `CODECOV_TOKEN` is already configured as a repo secret on this repo.

What changed

`pyproject.toml`

• `pytest-cov>=7.0` added to `[project.optional-dependencies] dev`
• `[tool.coverage.run] source = ["yt_dont_recommend"]`
• `[tool.coverage.report] show_missing = true, skip_empty = true`

`.github/workflows/ci.yml`

• `Run tests` step renamed to `Run tests with coverage`; invocation changed from `pytest tests/ -v` to `pytest --cov --cov-report=xml`
• New `Upload coverage to Codecov` step using `codecov/codecov-action@v5`. Gated on `matrix.os == 'ubuntu-latest'` so only one upload fires per CI run (matches mcp-clipboard's gating pattern, which keys off Python version — we don't matrix Python here, so OS is the natural single-axis gate). `fail_ci_if_error: false`, token via `secrets.CODECOV_TOKEN`.

`README.md`

• `Coverage` badge inserted between the `CI` and `Downloads` badges, matching mcp-clipboard's order.

Verification (local)

• `.venv/bin/pip install 'pytest-cov>=7.0'` then `pytest --cov --cov-report=xml` → 259/259 pass, `coverage.xml` generated (~105 KB)
• `coverage.xml` parsed cleanly by a spot check against Codecov's expected schema (it's a standard Cobertura-style XML)
• Baseline overall coverage ~44%. Breakdown:
  • `browser.py` 7% — Playwright-dependent, not unit-tested
  • `diagnostics.py` 0% — currently untested
  • `cli.py` 47%, `state.py` 63%, `config.py` 60%, `scheduler.py` 70%
  • `unblock.py` 83%, `blocklist.py` 86%, `clickbait.py` 87%
• Generated `coverage.xml` and `.coverage` cleaned up before commit. They are already ignored by pytest-cov / not checked in elsewhere.

Test plan (QA)

• `uv sync --extra dev` resolves with `pytest-cov>=7.0` pulled in
• `uv run pytest --cov --cov-report=xml` passes and writes `coverage.xml`
• After merge: first CI run on `main` uploads coverage to Codecov (watch the `Upload coverage to Codecov` step); Codecov badge on README starts resolving
• Coverage badge renders correctly on the PR / main README once data lands at Codecov

Out of scope

• Improving coverage on `browser.py` / `diagnostics.py` — a separate conversation; they require Playwright-level test infra.
• `codecov.yml` tuning (coverage thresholds, ignored paths) — mcp-clipboard doesn't have one either; defaults work until we have a reason to tighten them.

🤖 Generated with Claude Code

[codecov-commenter]

Welcome to Codecov 🎉

Once you merge this PR into your default branch, you're all set! Codecov will compare coverage reports and display results in all future pull requests.

ℹ️ You can also turn on project coverage checks and project coverage reporting on Pull Request comment

Thanks for integrating Codecov - We've got you covered ☂️

[cmeans]

[QA] Review — zero findings, Ready for QA Signoff.

What passed

Check	Result
uv sync --extra dev (local equivalent: pip install pytest-cov>=7.0)	Resolves cleanly
pytest --cov --cov-report=xml	259/259 pass; coverage.xml written (~105 KB, Cobertura schema). Overall line-rate ≈ 0.4368 — matches Dev's claimed ~44% baseline.
ruff check src/ tests/	All checks passed!
pyproject.toml additions	pytest-cov>=7.0 added to dev extras; [tool.coverage.run] and [tool.coverage.report] follow the standard coverage.py schema.
ci.yml diff	Run tests with coverage step invokes pytest --cov --cov-report=xml; new Upload coverage to Codecov step uses codecov/codecov-action@v5, fail_ci_if_error: false, token via secrets.CODECOV_TOKEN.
mcp-clipboard parity	mcp-clipboard gates on matrix.python-version == '3.13'; this PR gates on matrix.os == 'ubuntu-latest' because yt-dont-recommend doesn't matrix Python. Single upload per CI run — same intent via a different single-axis gate.
Coverage badge	[](https://codecov.io/gh/cmeans/yt-dont-recommend) — standard codecov.io schema; position between CI and Downloads matches mcp-clipboard's order.

Findings

None.

Verdict

Applying Ready for QA Signoff. The two post-merge checkboxes (Codecov upload on main + Coverage badge resolving) require data to land at Codecov and are genuinely unverifiable pre-merge — leaving them for the maintainer.

Heads-up: this PR and #22 both touch pyproject.toml (different sections — dev extras + [tool.coverage.*] here, license + authors + classifiers there). The second to merge will want a rebase. No merge-order labels applied — deferring that call to you.

[cmeans]

Dev action items (post-signoff):

• Decide merge order between this PR and chore: switch license to Apache-2.0 and fix PyPI metadata for badges #22. Both touch pyproject.toml in different sections (here: dev extras + [tool.coverage.*]; there: license / authors / classifiers) — no logical conflict, but the second to merge needs a rebase. Apply a merge-order: N label once you've picked.
• If this merges second: rebase onto main after chore: switch license to Apache-2.0 and fix PyPI metadata for badges #22 lands.
• Before the next release PR: ensure [Unreleased] in CHANGELOG.md gains an entry for this Codecov addition (Added). Tracked together with the pre-existing gap in docs(changelog): backfill [Unreleased] entries for PR #18, #20 (and #21 if merged) before next release #23.