Turbo test
Canonical Agent Skills library for Claude Code, Cursor, Codex, and other agents.From the repository description
npx -y skills add lgtm-hq/ai-skills --skill turbo-testAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
2.5 KB, 562 tokens by cl100k_base, as published. Nobody here has run it
Turbo-Themes Test Pipeline
Run the full build and test pipeline for turbo-themes. This project has a multi-stage pipeline where later stages depend on earlier outputs.
Do NOT use uv run lintro tst for this project — it does not cover the
full turbo-themes pipeline (examples, E2E, site build).
Full Pipeline (default)
When /turbo-test is invoked with no arguments, run all stages in order:
# 1. Lint (fast fail — catches formatting/type issues early)
uv run lintro chk
# 2. Build core + packages
bun run build
# 3. Build example projects (depends on core build output)
bun run examples:build
# 4. Unit tests (~2000 tests across ~74 files, with coverage)
bun run test
# 5. Example E2E tests (~139 tests across example projects)
bun run examples:test
If any stage fails, stop and fix before continuing to the next stage.
Subcommands
| Argument | Stages | Use when... |
|---|---|---|
| (none) | All 5 stages | Full validation (default) |
lint | Lint only | Quick formatting/style check |
build | Build + examples:build | Verifying build works |
unit | Unit tests only | Running tests after a build |
examples | examples:build + examples:test | Validating example projects |
quick | Lint + build + unit tests | Fast check, skip examples |
Optional E2E Stages
These are NOT part of the default pipeline. Run when explicitly needed:
# Smoke tests
bun run e2e:smoke
# Visual regression tests (snapshots generated on Linux CI only)
bun run e2e:visual
# Accessibility audits
bun run e2e:a11y
Note on visual regression: Snapshots are generated on Linux CI. Local
failures on macOS are expected. Use the maintenance-generate-snapshots.yml
workflow to regenerate snapshots after visual changes.
Rules
- Lint before build — catches issues early without waiting for compilation.
- Build before examples:build — examples depend on core package output.
- Build before test — some tests depend on build artifacts (tokens.json, CSS).
- Always include examples — example builds and tests are part of the standard pipeline, not optional extras. Do not skip them unless explicitly told to.
What ships with it
Read from the repository
Just SKILL.md. No reference files, no scripts.