cmp-qa-prep
Skill0 starsChanged 47 days agoMITView on GitHub
Bring up the E2E harness for a Kotlin/Compose Multiplatform app — boot the Android emulator, install the debug build, run the Maestro smoke — or, simpler, run the verify lane, which does all three itself. Use this when the user wants to run E2E/device tests on their CMP/KMP app, "prep my KMP test environment", "smoke test my Compose Multiplatform app", "run maestro on my CMP app", or "verify my CMP app runs on a device". Assumes the toolchain is installed (cmp-doctor) and an app exists (cmp-new).
- Pipes a download into a shell
In this repository17 other files it gives its agents.
- AGENTS.mdAGENTS.md
- template/AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- template/CLAUDE.mdCLAUDE.md
- llms.txtllms.txt
- npm-publishSkill
- plugin-refreshSkill
- cmp-auditSkill
- cmp-dev-clientSkill
- cmp-doctorSkill
- cmp-firebase-connectSkill
- cmp-inspectSkill
- cmp-newSkill
- cmp-previewSkill
- cmp-testSkill
- cmp-upgradeSkill
- grill-meSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 20
By hand, for an interactive session: check disk first (df -h ~, ≥ 3 GB free — Gradle fails with "No space left on device").
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
cmp-qa-prep — E2E harness bring-up for a CMP app
Current scaffolds ship Maestro flows (qa/e2e/*.yaml). Two ways to bring the harness up; the
first is almost always the one you want:
- Let the lane do it:
node qa/verify.mjs. With nothing attached it boots the project's AVD headless (CMP_AVDoverrides), installs the debug build, runs every flow inqa/e2e/, writes the receipt, and shuts the emulator down. This is the done-gate; a green run here IS the proof. By hand, for an interactive session: check disk first (
df -h ~, ≥ 3 GB free — Gradle fails with "No space left on device"); boot headless —emulator -avd <name> -no-window -no-audio -no-boot-anim -no-snapshot -gpu swiftshader_indirect— then
adb wait-for-device(the AVDcmp-doctorcreated; one AVD per app);./gradlew :composeApp:installDebug;maestro test qa/e2e/smoke.yaml(CLI:curl -fsSL https://get.maestro.mobile.dev | bash). Device proof is a checkpoint, never an inner loop — the project's PreToolUse hook reminds you.
Headless is the rule, not a preference. Those flags are HEADLESS_ARGS in the cmp profile's
device-provider.mjs, and they are what path 1 already uses — so do not hand-boot a windowed
emulator in front of a lane that would have booted its own. A window costs GPU and RAM the build
wants, makes the run depend on a desktop session, and differs from what CI does; -no-snapshot is
there so every lane boots the same cold device rather than inheriting yesterday's state. If a device
is already attached the lane uses it as-is, which is exactly how a windowed emulator quietly becomes
the thing your proof ran on (in create-cmp, node scripts/fleet-check.mjs says so when it finds
one).
Preconditions: toolchain present (JDK, Android SDK + AVD — cmp-doctor heals it); a project
scaffolded with the e2e toggle on (otherwise say so — there is no harness to bring up).
Legacy pre-Maestro scaffolds (qa/appium/, tests/appium/; the toggle was renamed from appium
in 0.3.0) are not covered here — migrate with cmp-upgrade, or see the skill's history.
Report
Disk headroom, emulator/AVD used, install exit code, and the smoke PASS/FAIL with the asserted
element. On failure, the concrete next step: low disk → clean Gradle caches; no AVD → cmp-doctor;
build error → surface the Gradle failure; device offline while adb devices is fine →
adb kill-server && adb start-server.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

