TypeScript Code Quality

This documentation was generated with the assistance of AI. Please report any inaccuracies.

Run this project’s TypeScript static-analysis and formatting tools — ESLint (flat config, plus Oxlint when it’s set up alongside ESLint), tsc --noEmit type-checking, Prettier --check, Biome when the project chose it, and Stylelint for CSS/SCSS when configured — then report every issue found (file, line, rule id, message), grouped by tool and classified by rule id. This is a read-only reporting skill: it runs the configured tools and reports results, it does not fix issues unless asked to as an explicit follow-up.

Purpose

iru-typescript-code-quality gives a single, framework-aware lint/type-check/format pass across the npm ecosystem’s fragmented tooling. It detects the project’s framework (npm/TypeScript library, React, Angular, React Native, Ionic, Hilla frontend, NestJS, Next.js) from package.json to pick the right invocation (e.g. ng lint vs a direct npx eslint) and the right rule-id prefixes to classify against, verifies each tool’s own config file actually exists before trusting a "zero issues" result from it, runs what the project’s own lint script runs, and — unlike Java’s Checkstyle/PMD/SpotBugs trio, which always analyze the whole project — every tool here except tsc accepts a glob directly, so a scoped invocation narrows what actually runs rather than just what gets reported.

Framework-specific behavior:

NestJS 12 scaffolds lint with Oxlint (oxlint --type-aware) plus Prettier and no ESLint, so Oxlint is the primary linter there; older Nest projects use ESLint plus Prettier like any other ESLint project.
Next.js 16 removed next lint, so the skill runs plain eslint with the eslint-config-next flat config (next lint only on Next 15 or older, or biome check when the project chose Biome), and runs next typegen before tsc --noEmit so the generated route types exist. Rule ids from @next/next/ and jsx-a11y/ are classified alongside the usual ones.
React Native resolves a toolchain first (expo in dependencies means expo, even with prebuild output; react-native without expo means metro, a bare CLI app). Metro apps use @react-native/eslint-config (the legacy .eslintrc.js with ESLint 8, or its /flat export; an ESLint 9 crash from an incompatible plugin is reported as a crash, never as zero issues), expo apps use eslint-config-expo. Both run as plain eslint over the whole app because expo lint only covers src/, app/ and components/. Prettier is scoped to source files (*/.{js,jsx,ts,tsx}) so the native android//ios/ projects are not flagged, and tsc --noEmit on Expo with TypeScript 6 may need "types": ["jest"], reported as a single configuration finding. Rule ids from react-native/, @react-native/, ft-flow/, expo/ and import/* are classified.

For UI projects (React/Vite, Next.js, Ionic, Hilla, Angular) it also sums the accessibility lint rules already in the ESLint buckets (jsx-a11y/*, angular-eslint’s template accessibility rules) into one accessibility subtotal so the accessibility level of the catalog’s UI contract (see Building user interfaces) can be read at a glance; when neither plugin is configured it says "accessibility lint not configured" instead, and never adds the plugin or edits the config.

Purpose group

TypeScript Development Tooling

Plugin

Shipped by the iru-typescript plugin of the ai-catalog marketplace (install with /plugin install iru-typescript@ai-catalog; see Plugins).

Invocation

/iru-typescript-code-quality
/iru-typescript-code-quality <path-or-glob>

Inputs

Argument Required Description Default

Argument	Required	Description	Default
`path-or-glob`	No	Scopes both the run and the report to specific files. ESLint, Oxlint, Prettier, and Stylelint accept the glob directly and only analyze what it matches; `tsc` has no path argument, so it always type-checks the whole project graph and the reported diagnostic lines are filtered afterward to ones whose file matches the scope.	None — omitting it checks and reports the whole project.

path-or-glob

Scopes both the run and the report to specific files. ESLint, Oxlint, Prettier, and Stylelint accept the glob directly and only analyze what it matches; tsc has no path argument, so it always type-checks the whole project graph and the reported diagnostic lines are filtered afterward to ones whose file matches the scope.

None — omitting it checks and reports the whole project.

Outputs

A plain-text summary in the conversation — never a written file.
Findings grouped first by tool (ESLint, Oxlint, Biome, tsc, Prettier, Stylelint), then by rule id/diagnostic code within each tool.
Per issue: file:line, rule id (or TS<code> or Biome lint/<group>/<rule> category), and message; Prettier findings are filed under a single formatting bucket since it has no rule ids.
A total issue count per tool and an overall total; a tool whose config isn’t found is reported as "not configured, skipped" rather than a misleading clean pass.
When the prompt supplies a baseline (a prior per-rule-id count, e.g. from iru-typescript-code-one-task-group), only the new-vs-baseline regressions per rule id — not the full current list.
For a UI project, an accessibility subtotal right below the overall total (or "accessibility lint not configured").
No source, ESLint, tsc, Prettier, or Stylelint configuration is ever modified.

Execution flow

flowchart TD A["Start /iru-typescript-code-quality\n(optional path-or-glob)"] --> B["Detect framework from package.json\n(library, react, angular, react-native [expo or metro], ionic, hilla-frontend, nestjs, nextjs)"] B --> C["Detect package manager\n(npm, pnpm, yarn) from lockfile/packageManager"] C --> D["Verify each tool's config exists\n(eslint.config.*/angular.json lint target,\ntsconfig.json, prettierrc, stylelintrc, oxlintrc)"] D --> D2["Pick linter from the lint script:\nOxlint for NestJS 12, plain eslint for Next 16,\nplain eslint for React Native (not expo lint),\nng lint for Angular"] D2 --> E{Tool configured?} E -- no --> F["Report '<tool> not configured, skipped'"] E -- yes --> G["Run ESLint (or ng lint) --format json,\nOxlint --format json if configured"] G --> H["Run tsc --noEmit --pretty false\n(-b for solution-style tsconfig;\nnext typegen first on Next.js)"] H --> I["Run Prettier --list-different"] I --> I2["Run Biome check\nif configured"] I2 --> J["Run Stylelint --formatter json\nif CSS/SCSS files + config exist"] J --> K["Read each generated report"] K --> L{Scope argument given?} L -- yes --> M["Filter tsc diagnostics to matching files\n(other tools already scoped by glob)"] L -- no --> N["Keep full project-wide results"] M --> O N --> O["Bucket issues by tool, then by rule id/TS code"] O --> P{Baseline supplied?} P -- yes --> Q["Report only new-vs-baseline regressions\nper rule id"] P -- no --> R["Report full per-tool, per-rule-id counts\nplus overall total"]

Dependencies

Invokes

None — iru-typescript-code-quality is a leaf skill; it only runs ESLint/Oxlint/tsc/Prettier/Stylelint and reads their generated reports, it does not call any other skill in this catalog.

Invoked by

Code — Step 3 and 7, quality baseline and comparison for TypeScript, NestJS and Next.js, via the iru-gate-runner agent.
TypeScript Code One Task Group — Steps 1 and 3.6, once per task group’s bucket, via the iru-gate-runner agent, to capture a pre-change baseline covering every file the bucket touches and later diff the post-change result against it.
NestJS Code One Task Group — Step 1 and 3.7, lint, tsc and Prettier baseline and regression check, via the iru-gate-runner agent.
Next.js Code One Task Group — Step 1 and 3.6, baseline and regression diff including typegen and tsc, via the iru-gate-runner agent.
React Native Code One Task Group — Step 1 baseline capture and the Step 3 quality regression check for React Native apps, via the iru-gate-runner agent.

Related agents

iru-gate-runner — both callers above always invoke this skill from inside a iru-gate-runner sub-agent they spawn, not inline.

Source

SKILL.md on GitHub — the file this page was generated from.