Skip to content
Agentmart
Submit a tool

run-quality-gates

Run the same lint, test, screenshot, and build checks as CI.

Skill by KotlinFoundation

IndexedMIT★ 28
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/kotlinfoundation-run-quality-gates

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Run the same lint, test, screenshot, and build checks as CI. Use before declaring any code change done, before commits/PRs, or when the user asks to validate/verify changes.

Use cases: testing, devops

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md and code. Checked Oct 26, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/run-quality-gates

What's inside

Skill name
run-quality-gates
Description
Run the same lint, test, screenshot, and build checks as CI. Use before declaring any code change done, before commits/PRs, or when the user asks to validate/verify changes.
License
MIT
Instructions
51 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# Run quality gates

All commands run from `MobileApp/`. Same gates as `.github/workflows/pr_checks.yml`, in order.

## 0. Environment keys (agent-visible signal)

The build defaults missing keys to `"testValue"`/empty and stays **green**, so a missing cloud key
never fails a gate. Run this first so you can SEE what's still a placeholder:

```bash
# Pass the current phase; omit --phase to check every phase.
./scripts/check_env.sh --phase <getting-started|integrations|publishing|monetization|growth>
```

It's non-breaking (exit 0). If it prints any `⚠️` line for a key **required by the current phase**,
**STOP and ask the developer** to supply that key before declaring the gates green — the compile
gates below will pass regardless.

```bash
# 1. Formatting / lint (auto-fix first, then verify)
./gradlew spotlessApply
./gradlew spotlessCheck

# 2. Unit + UI tests (commonTest runs in both suites)
./gradlew :shared:jvmTest :shared:testAndroidHostTest

# 3. Android debug build (also compiles :shared transitively)
./gradlew :androidApp:assembleDebug
```

Rules:

- **Changed UI?** These gates prove it compiles and behaves — they do **not** prove it *looks* right (screenshot tests aren't a PR gate). Use the **`verify-ui`** skill to render the screen to a PNG and check it.
- Do NOT run iOS builds/tests for routine validation — they are slow. Only when the change is iOS-specific or the user asks.
- Web check when web code changed: `./gradlew :shared:compileKotlinWasmJs :shared:compileKotlinJs`.
- Quick iOS-code compile check without a full build: `./gradlew :shared:compileAppleMainKotlinMetadata`.

## Do NOT (this is what causes a build loop)

The three scoped tasks above ARE the whole validation. Improvising around them is what spirals.

- **Never run `check`, `build`, or `clean build`.** They aggregate **every** target — including iOS — so an unrelated iOS cache/link failure fails the whole run and reads as "broken", tempting a retry. They're also far slower. Use the scoped gates, nothing wider.
- **Don't launch the app to "verify" a code change.** `:androidApp:assembleDebug` compiling green IS the Android gate. Do **not** auto-install + launch via `adb` and poll for the process — the launcher Activity is `.AppActivity` (Application class `.AndroidApp`), but detecting it via adb is fragile and is exactly what loops. To watch it render, use **`verify-ui`** (headless PNG) or ask the developer to hit Run in the IDE.
- **Run tasks never return** (`:desktopApp:run`, `:webApp:wasmJsBrowserDevelopmentRun`, `installDebug`+launch). A command that hasn't exited is **running, not hung** — start it once (in the background if you need the shell) and move on; do not kill and re-run.
- **A failed or slow command is not a retry signal.** Read the error → fix it, or STOP and report. Never re-run the same command hoping for a different result. The **first** build is slow (downloads JBR + Compose) — expected, not a hang.

---

*Cross-phase — every guide's validation steps call this skill.*

From KotlinFoundation/kmp-contest-starter-kit, MIT.

More from KotlinFoundation/kmp-contest-starter-kit

Signals

Stars
28
Forks
13
Open issues
1
Last commit (this folder)
Aug 10, 2026 (2 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from KotlinFoundation

Cite this page

run-quality-gates skill. Agent Mart. https://agentmart.dev/skills/kotlinfoundation-run-quality-gates (last updated 2026-10-24)