agentleFS
wrinfotelwrinfotel/verified-android-skills/skills/release-engineering/SKILL.md

release-engineering

Skill0 starsChanged 32 days agoMITView on GitHub

Android release readiness: the release build type must carry no debug tooling (isDebuggable) and never sign with the debug key, cleartext HTTP stays out of non-debug manifests, and intent-filtered components declare android:exported (Android 12+ install requirement). Use when touching build types, signing configs, or manifests. Core rules have machine checks in checks/.

8 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

7 rules. 0 are common, 2 are distinctive, and 5 are very specific to this project.

  • 0

    Common

    Most of the 7 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    After ANY build-type, signing, or manifest edit, run the Verification step (bottom of this file).

    line 16
  2. Very specific:

    MUST NOT set isDebuggable = true in the release block.

    line 21
  3. Very specific:

    MUST NOT point the release signingConfig at the debug key (signingConfigs.debug / getByName("debug")): the debug key is public knowledge …

    line 41
  4. Very specific:

    MUST NOT set android:usesCleartextTraffic="true" in the main (or release) manifest: "Starting with Android 9 (API level 28), cleartext …

    line 63
  5. Very specific:

    MUST declare android:exported on every activity, service, or broadcast receiver that has an <intent-filter>: "If your app targets Android …

    line 84
  6. Distinctive:

    Run after every build/manifest edit (from the project root):

    line 120
  7. Very specific:

    All four must pass (exit 0) before declaring the task done.

    line 129

“Common” is measured: kept by at least 6 in 10 of the 7 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Release Engineering

Operational rules for a coding agent. RIGHT/WRONG pairs are canonical. After ANY build-type, signing, or manifest edit, run the Verification step (bottom of this file).

1. No debug tooling in the release build type

MUST NOT set isDebuggable = true in the release block. The debug build type "includes debugging tools and is signed with the debug key"; the release build type is the distribution build ("applies ProGuard settings and is not signed by default"). A debuggable release ships the debugging surface to users and weakens app security.

WRONG:

buildTypes {
    release {
        isDebuggable = true
    }
}

RIGHT: leave the default (isDebuggable = false in release); do debug work in the debug build type.

2. Release never signs with the debug key

MUST NOT point the release signingConfig at the debug key (signingConfigs.debug / getByName("debug")): the debug key is public knowledge and "the debug build type ... is signed with the debug key". The release type "is not signed by default" — sign it explicitly with a release key (ideally via Play App Signing, see §5).

WRONG:

release {
    signingConfig = signingConfigs.getByName("debug")
}

RIGHT:

release {
    signingConfig = signingConfigs.getByName("release")
}

3. Cleartext HTTP stays out of non-debug manifests

MUST NOT set android:usesCleartextTraffic="true" in the main (or release) manifest: "Starting with Android 9 (API level 28), cleartext support is disabled by default", and enabling it app-wide is "this insecure configuration should be avoided whenever possible". If a domain genuinely needs HTTP, use a scoped domain-config in the network security config instead. The src/debug source set is exempt — pointing local dev tools at plain HTTP is normal there.

WRONG (main AndroidManifest.xml):

<application android:usesCleartextTraffic="true">

RIGHT:

<!-- src/debug/AndroidManifest.xml -->
<application android:usesCleartextTraffic="true" />

4. Intent-filtered components declare android:exported

MUST declare android:exported on every activity, service, or broadcast receiver that has an <intent-filter>: "If your app targets Android 12 or higher and contains activities, services, or broadcast receivers that use intent filters, you must explicitly declare the android:exported attribute" — otherwise "your app can't be installed on a device that runs Android 12 or higher." Launcher activity: exported true; most others: false.

WRONG:

<service android:name=".SyncService">
    <intent-filter>...</intent-filter>
</service>

RIGHT:

<service
    android:name=".SyncService"
    android:exported="false">
    <intent-filter>...</intent-filter>
</service>

5. What the checks cannot see (review-verified)

  • Versioning discipline: versionCode monotonic across releases, meaningful versionName.
  • Key management: Play App Signing (Google holds the app signing key; upload key rotates safely); keystores and passwords stay out of version control.
  • Shipping pipeline: App Bundles (AAB) for Play, R8 mapping file upload for deobfuscating crashes, staged rollouts, release notes.

Verification

Run after every build/manifest edit (from the project root):

bash <skill-dir>/checks/check_release_debuggable.sh  <project-or-module-dir>
bash <skill-dir>/checks/check_release_signing.sh     <project-or-module-dir>
bash <skill-dir>/checks/check_cleartext_traffic.sh   <project-or-module-dir>
bash <skill-dir>/checks/check_manifest_exported.sh   <project-or-module-dir>

All four must pass (exit 0) before declaring the task done. Checks are deterministic heuristics — they catch the documented anti-patterns above, not every release bug. §5 is verified by review against references/release.md.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.