| name | quality-standards |
| description | Use when choosing, interpreting, or reacting to a VilnaCRM website quality gate — what each gate enforces, the required make format then make lint order, and which specialist skill fixes a given failure. Triggers: "which gate", "make format failed", "make lint failed", "make lint-next / lint-tsc / lint-md / lint-deps / lint-metrics failed", "is this CI-green", and any temptation to lower a threshold in config/metrics-policy.json or to silence a finding with eslint-disable, @ts-ignore, @ts-nocheck, @ts-expect-error, prettier-ignore, or markdownlint-disable. Routes each failing gate to its specialist skill. |
Quality Standards
This is the catalog of the website's quality gates: what each one enforces, the
order to run them, the policy that protects them, and which specialist skill
fixes each failure. For the full run-before-PR orchestration see
../ci-workflow/SKILL.md; for the test-coverage
contract see the root agents.md.
Required Order
make format
make lint
make format runs Prettier only (no qlty fmt, no jscpd duplication pass —
neither is configured here). make lint is a verification gate, never a
mutating formatter. Run make format before any lint so Prettier output does not
re-trigger a lint failure. The complexity gate make lint-metrics is introduced by
issue #224 and is not yet on main; once #224 lands, run it as a separate hard gate
alongside make lint whenever you change src/ code.
Quality Gates
| Gate | Command |
|---|
| Format | make format |
| ESLint | make lint-next |
| TypeScript | make lint-tsc |
| Markdown | make lint-md |
| Architecture | make lint-deps |
| Metrics (planned, #224) | make lint-metrics |
| Aggregate | make lint |
make lint chains lint-next, lint-tsc, lint-md, and lint-deps in
sequence. make lint-metrics (rust-code-analysis, issue #224) is the complexity
gate and is not part of the make lint chain — treat it as an equal,
separate hard gate. See reference/gate-catalog.md
for what each gate enforces, its config file, and its common failures.
Protected Policy
These invariants make a gate meaningful. Breaking one to go green is not done.
- Do not lower a threshold or exclude a file in
config/metrics-policy.json. Reduce the
complexity instead (see complexity-management).
- Do not silence findings with
eslint-disable, @ts-ignore, @ts-nocheck,
@ts-expect-error, prettier-ignore, or markdownlint-disable. Fix the root
cause.
- Do not weaken a rule in
eslint.config.mjs, .dependency-cruiser.js,
tsconfig.json, or .markdownlint.yaml to make a change pass.
- Do not commit regenerated Playwright visual snapshots unless the visual change
is intentional; regenerate with
make test-visual-update and review the diff.
- Reuse Makefile targets — never invoke the underlying binary with relaxed flags
to bypass a gate.
Route Each Failure To Its Specialist
When a gate fails, hand the fix to the skill that owns that concern:
make lint-next / make lint-tsc / make lint-md (ESLint, TypeScript, markdownlint)
— ../frontend-quality-workflow/SKILL.md owns these
gates; it routes a component, hook, form, or typing fix to
../frontend-component-development/SKILL.md, file
placement or layering to ../code-organization/SKILL.md, and
doc drift to ../documentation-sync/SKILL.md.make lint-deps (dependency-cruiser: features-import-via-public-api,
no-cross-feature-imports, feature-allowed-folders,
src-feature-name-kebab-case): ../architecture/SKILL.md
and ../code-organization/SKILL.md.make lint-metrics (cyclomatic, cognitive, ABC, NARGS, NEXITS, Halstead,
MI, file-size budgets): ../complexity-management/SKILL.md.- New or updated docs (feature READMEs, agent guides, skill files):
../documentation-creation/SKILL.md.
make test-visual (visual regression) and Lighthouse budgets:
../figma-design-check/SKILL.md and
../frontend-performance-accessibility/SKILL.md.- Jest / Playwright e2e / Stryker coverage and assertion gaps: the root
agents.md test-scenario policy.
- PR review comments from
make pr-comments: ../code-review/SKILL.md.
Focused Test Gates
Run the suite that actually exercises the change; one change often needs more
than one. The root agents.md policy governs scenario coverage.
| Change type | Command |
|---|
| Component or hook | make test-unit-client |
| Apollo server logic | make test-unit-server |
| Both unit layers | make test-unit-all |
| User-facing journey | make test-e2e |
| Rendered UI or styling | make test-visual |
| Performance or a11y | make lighthouse-mobile |
| Mutation strength | make test-mutation |
| Traffic or load | make test-load |
| Memory leaks | make test-memory-leak |
Any unit suite runs locally without Docker when prefixed with CI=1 (for
example CI=1 make test-unit-all).
Verification
make format
make lint
make lint-metrics
Then run the focused test gate(s) the change touches. Work is not done until the
relevant commands have actually been run and pass.
Related Guides
Before applying this skill, confirm the active task against
../AI-AGENT-GUIDE.md and
../SKILL-DECISION-GUIDE.md so every relevant skill
is consulted. The root agents.md test-coverage policy applies to any behavior a
change touches.
Line Length Disclosure
Prettier printWidth is 100. Before presenting changes, check changed text files
for lines longer than 100 characters. If any exist, tell the user each
path:line and the measured character count. Treat this as disclosure, not
failure, unless a project gate fails.
Supporting Files
