KMP Code Quality

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

Run whichever static-analysis tools a Kotlin Multiplatform (Android + iOS) project actually has wired — detekt, ktlint, Android Lint and SwiftLint — over the shared module’s every source set, the Android app module and optionally the Swift host app, and report every issue they find. This is a report-only skill: it never runs formatters or baseline generators, never fixes issues and never changes or installs tool configuration.

Purpose

iru-kmp-code-quality first verifies which tools are really wired, because KMP breaks the usual assumptions: a plain detekt 1.x detekt task analyzes nothing on a KMP module unless source is configured (reported as "wired but not covering KMP source sets", never as clean), detekt 2.x uses per-source-set tasks, ktlint checks one task per source set, the AGP KMP library plugin registers no Lint task unless com.android.lint is applied, and SwiftLint needs both iosApp/.swiftlint.yml and an installed binary. It then runs every wired task in one Gradle invocation with --continue, locates and parses the Checkstyle XML, ktlint plain reports, Android Lint XML and SwiftLint JSON into one per-tool, per-rule-id summary, and, when the prompt supplies a baseline, reports only what is new (matched on tool + rule id + file). It is the KMP equivalent of Android Code Quality, Kotlin Code Quality and Swift Code Quality.

Purpose group

Kotlin Multiplatform Development Tooling

Plugin

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

Invocation

/iru-kmp-code-quality
/iru-kmp-code-quality <scope>
/iru-kmp-code-quality <scope>
module: <gradle path>
tools: detekt,ktlint,lint,swiftlint
baseline: <prior per-tool / per-rule-id counts or itemized list>

Inputs

Argument Required Description Default

<scope>

No

A file path, class/file simple name, module (:shared, androidApp) or source-set name (commonMain, iosMain, androidMain). It narrows which tasks run when it maps to a module or source set and filters the parsed report, since none of the tools take a per-file selector.

None — the whole project.

module

No

Restrict the run to one Gradle module.

All modules

tools

No

Comma-separated subset of detekt, ktlint, lint, swiftlint.

Every wired tool

baseline

No

A prior per-tool, per-rule-id count or itemized list from an earlier run; everything after baseline: is treated as that prior result and only new issues are reported.

None — report plain counts.

Tool wiring checks (Step 1):

Tool How it is detected and run

detekt

Generation from the plugin id (io.gitlab.arturbosch.detekt is 1.x, dev.detekt is 2.x). 1.x: per-compilation tasks (detektMetadataCommonMain, detektAndroidMain, …​) or the plain detekt task only when source is configured; 2.x: detekt<Set>SourceSet tasks.

ktlint

ktlintCheck (one check per source set plus ktlintKotlinScriptCheck), or Spotless kotlin { ktlint() }.

Android Lint

lintDebug on the Android app module; on the KMP module lint only when com.android.lint is applied (or lintDebug with androidTarget()).

SwiftLint

Only when iosApp/.swiftlint.yml exists and the swiftlint binary is installed (installation offered, never done without asking); not applicable off macOS.

Outputs

  • A plain-text report in the conversation — never a written file.

  • An overall issue total, then per tool: ran / not wired / wired but not covering KMP source sets / not installed, the governing config, baseline or .editorconfig, the issue count, severity breakdown and one line per rule bucket (detekt rule, ktlint ruleset:rule-id, Lint issue id, SwiftLint rule_id), grouped by source set.

  • The file:line list when the scope is smaller than the whole project.

  • With a baseline: only the new-vs-baseline regressions, with a note when a tool was newly wired or the detekt generation changed.

  • A "No Kotlin/Swift static analysis configured" message when nothing is wired.

  • Nothing is fixed, formatted or reconfigured.

Execution flow

flowchart TD A["Start /iru-kmp-code-quality\n(scope, tools, baseline)"] --> B["Step 0: resolve scope, module,\ntools, baseline;\ncheck JDK / Android SDK / gradlew"] B --> C["Step 1: verify which tools are wired\n(detekt generation, ktlint, Lint,\nSwiftLint) from tasks --all"] C --> D{"Any tool wired?"} D -- no --> Z["Stop: 'No Kotlin/Swift static\nanalysis configured'"] D -- yes --> E["Step 2: run all wired Gradle tasks\nwith --continue (exit codes are information)"] E --> F{"SwiftLint wired?"} F -- yes --> G["Run swiftlint lint\n--reporter json"] F -- no --> H["Step 3: locate reports by listing\nbuild/reports"] G --> H H --> I{"Tool failed with a compile or\nresolution error?"} I -- yes --> J["Report tool failure,\ndo not read stale reports"] I -- no --> K["Parse detekt / ktlint / Lint /\nSwiftLint reports, de-duplicate"] J --> L K --> L["Step 4: bucket by tool, rule id\nand source set"] L --> M{"Baseline supplied?"} M -- yes --> N["Report only new issues\n(tool + rule id + file)"] M -- no --> O["Report totals and\nper-rule counts"] N --> P["Done: no fixes, no config changes"] O --> P

Dependencies

Invokes

None — iru-kmp-code-quality is a leaf skill; it only runs Gradle tasks, swiftlint and inline python3 parsing, and it does not call any other skill in this catalog. Its prerequisite checks and module detection follow KMP Test by reference only.

Invoked by

  • KMP Code One Task Group — Step 1 (captures the bucket’s pre-change quality baseline) and Step 3.6 (diffs the post-change result against that baseline), through the iru-gate-runner agent.

  • Code — Step 3 (project-wide quality baseline per language key) and Step 7 (final comparison against that baseline), resolving the kmp key to this skill and running it through the iru-gate-runner agent.

Related agents

  • iru-gate-runner — both callers always run this skill from inside an iru-gate-runner sub-agent, so only the compact per-rule summary reaches their context.

Source

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