Android Permissions Security logo

Android Permissions Security

OrganizationPopular
android
android-permissions-security

Audits, detects gaps, and remediates Android permissions and IPC component security vulnerabilities. Use when reviewing AndroidManifest.xml, defining custom permissions, securing Services, Receivers, and Providers, verifying caller signatures and UID using Binder, or implementing runtime permission flows.

Overview

Publisherandroid
Repositoryskills
Skill nameandroid-permissions-security
Stars
7.8K
Forks
512
Bundled files
3
LicenseApache-2.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 3 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by android on GitHub. Read the source before you install it.

Installation

Install the Android Permissions Security AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/android/skills/tree/main/security/android-permissions-security
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/android/skills.git /tmp/skills
mkdir -p .claude/skills
cp -r /tmp/skills/security/android-permissions-security .claude/skills/android-permissions-security
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Android Permissions Security in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Android Permissions Security on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Android Permissions Security is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Guidelines for auditing permission vulnerabilities, enforcing least privilege, securing IPC boundaries, and handling runtime permissions across the Android ecosystem.

Glossary

  • Runtime Permission: Dangerous permission (e.g., Camera, Location) that must be granted by the user at runtime on Android 6.0+ (API 23+).
  • Custom Permission: A permission defined by an app (<permission>) to restrict access to its exported components.
  • Protection Level: Attribute defining how the OS grants the permission (normal, dangerous, signature, knownSigner).
  • knownSigner: Protection flag (API 31+) granting signature-level access to partner apps matching trusted certificate digests.
  • SigningInfo: Object (API 28+) providing signing history and multiple signer certificates for an application.
  • Binder Calling Identity: Reliable UID representing the calling process obtained using Binder.getCallingUid().
  • URI Permission: Temporary, granular access granted to a content URI using grantUriPermissions and intent flags.

Prerequisites

  • Add androidx.core:core-ktx and androidx.activity:activity-ktx (or androidx.fragment:fragment-ktx) dependencies to the app-level build.gradle file for ContextCompat and ActivityResultContracts.
  • For knownSigner partner permissions (android:knownCerts), ensure the project compileSdk and targetSdk are at least 31 (Android 12).
  • For broadcast identity sharing (BroadcastOptions.setShareIdentityEnabled), ensure compileSdk and targetSdk are at least 34 (Android 14).

1. Security audit and permission gap detection playbook

When auditing an Android project for permission and IPC security vulnerabilities, systematically inspect each area:

Vulnerability AreaSearch In / PatternRiskRemediation Action
Weak Custom Permission ProtectionAndroidManifest.xml with android:protectionLevel="normal", "dangerous", or "signatureOrSystem"Unauthorized apps request or claim permission without restrictionUpgrade to signature or `signature
Overly Broad URI Permission Grants<grant-uri-permission android:pathPrefix="/" /> or pathPrefix="" in <provider>Leaks all provider data to third partiesReplace with explicit scoped subpath (e.g. android:pathPrefix="/shared/"). Retain android:grantUriPermissions="true". Remove any wildcard pathPrefix="/".
Missing ContentProvider Split Permissions<provider> with only generic android:permission or missing android:readPermission and android:writePermissionRead-only callers execute write operations or write-only callers access read operationsConfigure granular android:readPermission="...READ_DATA" and android:writePermission="...WRITE_DATA", AND declare the custom <permission> tags with android:protectionLevel="signature" (or `signature
Spoofable Caller Identity Checksintent.getStringExtra("calling_package"), intent.getStringExtra("sender"), callingPackage == "..." in ServicesMalicious callers forge intent extras or stringsRemove string checks; authenticate caller UID cryptographically using Binder.getCallingUid() and pm.hasSigningCertificate().
Bound Service Permission BypasscheckCallingOrSelfPermission() or enforceCallingOrSelfPermission() in Binder AIDL stubsFalls back to host app UID outside IPC or when identity is cleared, causing Confused Deputy bypassReplace with enforceCallingPermission("...WRITE_DATA", ...) or checkCallingPermission("...WRITE_DATA").
Missing Method-Level Checks in Bound ServicesBound Service AIDL methods mutating state without enforcing write permissionsCallers with read-only binding access invoke mutating write operationsAdd enforceCallingPermission("...WRITE_DATA", "Caller lacks WRITE_DATA permission") inside mutating AIDL methods.
Flawed Runtime Permission FlowMissing shouldShowRequestPermissionRationale(), missing permanent denial handling, or re-requesting in loopsPoor user experience, infinite loops, or inability to recover permissionsImplement 3-state flow: check grant -> check rationale dialog -> launch contract. On permanent denial, redirect to Settings.ACTION_APPLICATION_DETAILS_SETTINGS.
Simultaneous Location RequestsarrayOf(ACCESS_FINE_LOCATION, ACCESS_BACKGROUND_LOCATION) in one requestRejected by Android 11+ (API 30+)Request foreground location (ACCESS_FINE_LOCATION, ACCESS_COARSE_LOCATION) first. Request ACCESS_BACKGROUND_LOCATION only after foreground is granted in a separate user step.
Broad Storage Permissions for MediaRequesting READ_EXTERNAL_STORAGE or READ_MEDIA_IMAGES for user image selectionOver-privilege, user privacy violationMigrate to zero-permission Photo Picker (ActivityResultContracts.PickVisualMedia).
Cached Permission StateStoring permission status in var cachedLocationPermissionGranted = ...Fails when permission is revoked or expiresQuery ContextCompat.checkSelfPermission() dynamically at invocation time and wrap protected calls in try-catch blocks catching SecurityException.
Insecure BroadcastssendBroadcast(intent) without package target or receiver permission; receiver without sender authenticationData interception or command spoofingSet explicit package (intent.setPackage(...)), pass receiver permission, and enable BroadcastOptions.setShareIdentityEnabled(true) (API 34+) to verify sentFromPackage.

2. Custom permissions and protection levels

When declaring custom permissions or partner certificates in the manifest, you must follow the instructions in the <permission> element guide.

  • signature: Restricts access to apps signed with the same developer key.
  • signature|knownSigner (API 31+): Restricts access to partner apps whose signing certificate digests match declared hashes in android:knownCerts.
  • normal or dangerous : Do not use for private inter-app IPC.
  • signatureOrSystem : Deprecated (API 23). Upgrade deprecated signatureOrSystem permissions to signature. Avoid in non-system apps.

Manifest snippets

Signature-level custom permission

Use signature permissions when restricting component access exclusively to apps signed with the same developer certificate by declaring the <permission> in the manifest and applying android:permission to the target component.

xml
<permission
    android:name="com.example.snippets.permission.ACCESS_SECURE_API"
    android:protectionLevel="signature" />
KnownSigner partner permission (API 31+)

Use signature|knownSigner permissions when granting access to specific external partner apps by declaring the <permission> in the manifest with android:knownCerts referencing a resource array of trusted SHA-256 certificate digests.

xml
<resources xmlns:tools="http://schemas.android.com/tools">
    <!-- SHA-256 fingerprints of trusted partner certificates (lowercase, no colons) -->
    <string-array name="trusted_partner_certs" tools:ignore="Typos">
        <item>103938ee4537e59e8ee792f654504fb8346fc6b346d0bbc4415fc339fcfc8ec1</item>
    </string-array>
</resources>
xml
<permission
    android:name="com.example.snippets.permission.PARTNER_API"
    android:protectionLevel="signature|knownSigner"
    android:knownCerts="@array/trusted_partner_certs" />

3. Exported component protection and URI permissions

  • Private components : Always set android:exported="false".
  • Exported components : Require permissions (android:permission, android:readPermission, android:writePermission).
  • Granular URI Grants : Retain android:grantUriPermissions="true" for secure sharing, but scope grants to specific sub-paths. Never use wildcard pathPrefix="/".
  • Custom Permission Declarations : When splitting read and write permissions on a provider, you MUST declare the <permission> tags with android:protectionLevel="signature" (or signature|knownSigner) at the <manifest> root level.
  • Remediating Existing Wildcard Grants : You MUST find and remove any existing <grant-uri-permission android:pathPrefix="/" /> or pathPrefix="" and replace it with an explicit scoped subpath (e.g. android:pathPrefix="/shared/").
  • Granular UriPermission Management : Instead of broad manifest grants, use runtime android.content.UriPermission tokens for least-privilege per-URI sharing:
    • Grant temporary access : context.grantUriPermission(targetPackage, uri, Intent.FLAG_GRANT_READ_URI_PERMISSION) or attach Intent.FLAG_GRANT_READ_URI_PERMISSION to outgoing intents.
    • Persist access : The receiving app calls contentResolver.takePersistableUriPermission(uri, modeFlags) for durable access across reboots.
    • Audit active grants : Inspect contentResolver.persistedUriPermissions (List<UriPermission>) to verify granted URIs (uriPermission.uri, uriPermission.isReadPermission(), uriPermission.isWritePermission()).
    • Revoke access : Call context.revokeUriPermission(uri, modeFlags) or contentResolver.releasePersistableUriPermission(uri, modeFlags) as soon as operations conclude.

Manifest snippet: Complete secure ContentProvider declaration

Declare the split read and write permissions at the <manifest> root:

xml
<permission
    android:name="com.example.snippets.permission.READ_DATA"
    android:protectionLevel="signature" />
<permission
    android:name="com.example.snippets.permission.WRITE_DATA"
    android:protectionLevel="signature" />

Then configure the provider with those permissions and a scoped URI grant:

xml
<provider
    android:name=".permissions.SecureDataProvider"
    android:authorities="com.example.snippets.partnerprovider"
    android:exported="true"
    android:readPermission="com.example.snippets.permission.READ_DATA"
    android:writePermission="com.example.snippets.permission.WRITE_DATA"
    android:grantUriPermissions="true">
    <grant-uri-permission android:pathPrefix="/shared/" />
</provider>

Runtime granular UriPermission management snippet

kotlin
object UriPermissionManager {

    /** Grants temporary read access to a specific URI for a partner package. */
    fun grantScopedUriAccess(context: Context, targetPackage: String, uri: Uri) {
        context.grantUriPermission(
            targetPackage,
            uri,
            Intent.FLAG_GRANT_READ_URI_PERMISSION
        )
    }

    /** Audits active persisted URI permissions held by the application. */
    fun getActivePersistedPermissions(context: Context): List<UriPermission> {
        return context.contentResolver.persistedUriPermissions
    }

    /** Revokes temporary URI access when data transfer completes. */
    fun revokeScopedUriAccess(context: Context, uri: Uri) {
        context.revokeUriPermission(
            uri,
            Intent.FLAG_GRANT_READ_URI_PERMISSION or Intent.FLAG_GRANT_WRITE_URI_PERMISSION
        )
    }
}

4. Cryptographic caller identity verification

Never trust caller package names provided in Intents or string arguments (getCallingPackage(), intent.getStringExtra("calling_package"), intent.getStringExtra("sender"), intent.getStringExtra("package")). Intent extras and string package names are forged by malicious applications.

Critical vulnerability: Package name spoofing using intent extras

kotlin
// VULNERABLE PATTERN: DO NOT DO THIS
val callingPackage = intent?.getStringExtra("calling_package")
if (callingPackage != null && SignatureUtils.verifyPartnerPackage(this, callingPackage)) {
    // A malicious app passes "com.example.partner" in the extra.
    // The signature check verifies the installed partner on disk, but the CALLER was malicious!
}

Secure pattern: Authenticate calling UID cryptographically

  1. Completely remove all intent.getStringExtra("calling_package"), intent.getStringExtra("sender"), and callingPackage == "..." checks.
  2. Obtain caller identity using Binder.getCallingUid().
  3. Verify caller certificate using PackageManager.hasSigningCertificate() (API 28+) or PackageManager.getPackagesForUid().

Complete CallerVerifier implementation

Apply this to the service that currently performs the spoofable check.

  • Placement : declare CallerVerifier in the same file as the service you are fixing (for example inside SecureDataService.kt). Do not move the verification into a new file and leave the original service untouched.
  • Nullability : packageInfo.signatures is Array<Signature>?. Copy the ?: return false guard exactly as written below. Iterating packageInfo.signatures directly fails to compile with "Non-nullable value required to call an 'iterator()' method in a for-loop."
kotlin
object CallerVerifier {
    private const val TRUSTED_PARTNER_SHA256 =
        "A1B2C3D4E5F60708090A0B0C0D0E0F1011121314151617181920212223242526"

    fun isCallerAuthorized(context: Context): Boolean {
        val callingUid = Binder.getCallingUid()
        if (callingUid == android.os.Process.myUid()) return true

        val pm = context.packageManager
        // Modern API 28+ check by UID:
        if (Build.VERSION.SDK_INT >= Build.VERSION_CODES.P) {
            val certBytes = hexStringToByteArray(TRUSTED_PARTNER_SHA256)
            if (pm.hasSigningCertificate(callingUid, certBytes, PackageManager.CERT_INPUT_SHA256)) {
                return true
            }
        }

        // Fallback for legacy APIs:
        val callingPackages = pm.getPackagesForUid(callingUid) ?: return false
        for (pkg in callingPackages) {
            if (verifyPackageSignature(pm, pkg)) {
                return true
            }
        }
        return false
    }

    @Suppress("DEPRECATION")
    private fun verifyPackageSignature(pm: PackageManager, packageName: String): Boolean {
        return try {
            val packageInfo = pm.getPackageInfo(packageName, PackageManager.GET_SIGNATURES)
            val signatures = packageInfo.signatures ?: return false
            for (sig in signatures) {
                val digest = java.security.MessageDigest.getInstance("SHA-256").digest(sig.toByteArray())
                val hex = digest.joinToString("") { "%02X".format(it) }
                if (hex.equals(TRUSTED_PARTNER_SHA256, ignoreCase = true)) return true
            }
            false
        } catch (e: PackageManager.NameNotFoundException) {
            false
        }
    }

    private fun hexStringToByteArray(s: String): ByteArray {
        val len = s.length
        val data = ByteArray(len / 2)
        for (i in 0 until len step 2) {
            data[i / 2] = ((Character.digit(s[i], 16) shl 4) + Character.digit(s[i + 1], 16)).toByte()
        }
        return data
    }
}

5. Method-level Binder enforcement in bound services (SecureBoundService.kt)

In bound services exposing AIDL Binder stubs, protect state-modifying methods using enforceCallingPermission().

Critical security rule: Never use CallingOrSelfPermission in services

  • NEVER use checkCallingOrSelfPermission() or enforceCallingOrSelfPermission().
  • Why: During an active Binder transaction, this method checks the caller's permission. However, if the method is invoked locally (outside of an active IPC) or if the IPC calling identity is dropped, it falls back to checking the host app's permissions. Because the host app holds the permission, this fallback grants access and bypasses the security boundary, causing a Confused Deputy vulnerability.
  • MUST USE: enforceCallingPermission("com.example.snippets.permission.WRITE_DATA", "Caller lacks WRITE_DATA permission") or checkCallingPermission("com.example.snippets.permission.WRITE_DATA").

Complete SecureBoundService implementation

kotlin
class SecureBoundService : Service() {

    private val binder = object : IMyService.Stub() {
        override fun getData(): String {
            // Read-only operation guarded by manifest-level permission
            return "Confidential Data"
        }

        override fun modifyData(newData: String) {
            // MUST use enforceCallingPermission or checkCallingPermission.
            // NEVER use checkCallingOrSelfPermission or enforceCallingOrSelfPermission.
            this@SecureBoundService.enforceCallingPermission(
                "com.example.snippets.permission.WRITE_DATA",
                "Caller lacks WRITE_DATA permission"
            )
            updateInternalState(newData)
            Log.d("SecureBoundService", "Data modified to: $newData with proper WRITE_DATA permission check")
        }
    }

    override fun onBind(intent: Intent?): IBinder = binder

    private fun updateInternalState(data: String) {
        // Internal state update logic
    }
}

6. Safe runtime permission workflow (RuntimePermissionsActivity.kt)

When implementing runtime permission requests or handling permission denial states, you must follow the instructions in the Request runtime permissions guide and Permissions overview. Implement a 3-state flow without caching states or creating infinite retry loops:

  1. Check Granted State : ContextCompat.checkSelfPermission(this, permission) == PackageManager.PERMISSION_GRANTED.
  2. Check Rationale : shouldShowRequestPermissionRationale(permission) to present educational context before requesting.
  3. Launch Contract : ActivityResultContracts.RequestPermission().
  4. Handle Permanent Denial : In the callback, if permission was denied and !shouldShowRequestPermissionRationale(permission), guide the user to Settings.ACTION_APPLICATION_DETAILS_SETTINGS.

Complete RuntimePermissionsActivity.kt implementation

kotlin
class RuntimePermissionsActivity : ComponentActivity() {

    private val cameraLauncher =
        registerForActivityResult(ActivityResultContracts.RequestPermission()) { isGranted ->
            if (isGranted) {
                startCameraPreview()
            } else {
                if (!shouldShowRequestPermissionRationale(Manifest.permission.CAMERA)) {
                    // User selected 'Don't ask again' or permanently denied.
                    // Direct user to Application Details Settings.
                    val intent = Intent(Settings.ACTION_APPLICATION_DETAILS_SETTINGS).apply {
                        data = Uri.fromParts("package", packageName, null)
                    }
                    startActivity(intent)
                } else {
                    showSnackbar("Camera permission is required to preview camera feed.")
                }
            }
        }

    fun requestCameraPermissionSafely() {
        val permission = Manifest.permission.CAMERA
        when {
            ContextCompat.checkSelfPermission(this, permission) == PackageManager.PERMISSION_GRANTED -> {
                startCameraPreview()
            }
            shouldShowRequestPermissionRationale(permission) -> {
                showSnackbar("Camera permission is needed to preview camera feed.")
                cameraLauncher.launch(permission)
            }
            else -> {
                cameraLauncher.launch(permission)
            }
        }
    }

    private fun startCameraPreview() {
        Log.d("RuntimePermissionsActivity", "Camera preview started")
    }

    private fun showSnackbar(msg: String) {
        Log.i("RuntimePermissionsActivity", msg)
    }
}

7. Zero-permission alternatives and sequential flows

When implementing photo or media selection, you must follow the instructions in the Photo Picker API guide to avoid requesting broad storage permissions.

  • Photo Picker : Prefer ActivityResultContracts.PickVisualMedia() over requesting wide storage permissions (READ_MEDIA_IMAGES, READ_EXTERNAL_STORAGE).
  • Sequential Location Flow : Request ACCESS_FINE_LOCATION and ACCESS_COARSE_LOCATION first. Never request ACCESS_BACKGROUND_LOCATION simultaneously. Only request background location in a separate step after foreground is granted.

Photo picker implementation

kotlin
private val photoPickerLauncher =
    registerForActivityResult(ActivityResultContracts.PickVisualMedia()) { uri ->
        if (uri != null) {
            handleImageUri(uri)
        }
    }

fun selectPhoto() {
    photoPickerLauncher.launch(
        PickVisualMediaRequest(ActivityResultContracts.PickVisualMedia.ImageOnly)
    )
}

Sequential location request implementation

kotlin
private val foregroundLocationLauncher =
    registerForActivityResult(ActivityResultContracts.RequestMultiplePermissions()) { permissions ->
        val fineGranted = permissions[Manifest.permission.ACCESS_FINE_LOCATION] ?: false
        val coarseGranted = permissions[Manifest.permission.ACCESS_COARSE_LOCATION] ?: false
        if (fineGranted || coarseGranted) {
            startForegroundLocationUpdates()
        }
    }

fun requestForegroundLocation() {
    foregroundLocationLauncher.launch(
        arrayOf(
            Manifest.permission.ACCESS_FINE_LOCATION,
            Manifest.permission.ACCESS_COARSE_LOCATION
        )
    )
}

// Background location requested only AFTER foreground is granted and user explicitly opts in:
private val backgroundLocationLauncher =
    registerForActivityResult(ActivityResultContracts.RequestPermission()) { isGranted ->
        if (isGranted) {
            startBackgroundTracking()
        }
    }

8. Stateless checks and error boundaries

  • Never cache permission states in in-memory variables (e.g. cachedLocationPermissionGranted). Check ContextCompat.checkSelfPermission() dynamically before each operation.
  • Wrap API calls in try-catch : Catch SecurityException around permission-protected framework calls (e.g., location queries) to handle one-time permission revocation gracefully.
kotlin
fun performLocationAccess() {
    val fine = ContextCompat.checkSelfPermission(this, Manifest.permission.ACCESS_FINE_LOCATION)
    val coarse = ContextCompat.checkSelfPermission(this, Manifest.permission.ACCESS_COARSE_LOCATION)
    if (fine != PackageManager.PERMISSION_GRANTED && coarse != PackageManager.PERMISSION_GRANTED) {
        requestForegroundLocation()
        return
    }

    val provider = if (fine == PackageManager.PERMISSION_GRANTED) {
        LocationManager.GPS_PROVIDER
    } else {
        LocationManager.NETWORK_PROVIDER
    }

    try {
        val location = locationManager.getLastKnownLocation(provider)
        processLocation(location)
    } catch (e: SecurityException) {
        Log.e("LocationAccess", "Permission revoked at runtime", e)
    }
}

9. Secure broadcasts (API 34+ share identity)

Use explicit intents, required receiver permissions, or Android 14+ broadcast options:

kotlin
// Sender: Enforce permission and share identity
val intent = Intent("com.example.snippets.permission.ACTION_SECRET_UPDATE").apply {
    setPackage("com.example.partner") // Explicit target
}
val isUdc = Build.VERSION.SDK_INT >= Build.VERSION_CODES.UPSIDE_DOWN_CAKE
val options = if (isUdc) {
    BroadcastOptions.makeBasic().apply {
        setShareIdentityEnabled(true)
    }.toBundle()
} else null

sendBroadcast(intent, "com.example.snippets.permission.RECEIVE_SECRET_UPDATE", options)
kotlin
// Receiver: Validate sender on Android 14+
class ProtectedReceiver : BroadcastReceiver() {
    override fun onReceive(context: Context, intent: Intent) {
        if (intent.action == "com.example.snippets.permission.ACTION_SECRET_UPDATE") {
            if (Build.VERSION.SDK_INT >= Build.VERSION_CODES.UPSIDE_DOWN_CAKE) {
                val sender = sentFromPackage
                if (sender != "com.example.trusted_sender") {
                    return // Reject unauthorized sender
                }
            }
            processUpdate(intent)
        }
    }

    private fun processUpdate(intent: Intent) {}
}

Legacy compatibility (Pre-API 34) : For apps targeting API 33 or lower, attach an immutable PendingIntent extra (FLAG_IMMUTABLE) to the broadcast and authenticate the sender using pendingIntent.creatorPackage.


Quality checklist

DO

  • MUST declare <permission android:name="..." android:protectionLevel="signature" /> (or signature|knownSigner) at the <manifest> root for custom permissions and ContentProvider split permissions.
  • MUST explicitly set android:exported="false" for all private components to remove the need for custom permission management.
  • MUST authenticate callers cryptographically using Binder.getCallingUid() and certificate hashes (PackageManager.hasSigningCertificate or getPackagesForUid) rather than checking spoofable intent extras.
  • MUST check shouldShowRequestPermissionRationale() before runtime prompts and during denial handling.
  • MUST direct user to Settings.ACTION_APPLICATION_DETAILS_SETTINGS when permission is permanently denied.
  • MUST enforce method-level checks with enforceCallingPermission() or checkCallingPermission() on modifying AIDL methods.
  • MUST scope ContentProvider URI permissions to explicit sub-paths (e.g. android:pathPrefix="/shared/") while retaining android:grantUriPermissions="true".
  • MUST catch SecurityException around protected API calls.
  • MUST request foreground and background location permissions sequentially, never simultaneously.

NEVER

  • NEVER extract or verify package names from Intent extras (intent.getStringExtra("calling_package"), intent.getStringExtra("sender"), intent.getStringExtra("package")). Always authenticate using Binder.getCallingUid() and certificates.
  • NEVER remediate caller spoofing by merely adding manifest permissions while keeping spoofable intent extra checks.
  • NEVER use checkCallingOrSelfPermission() or enforceCallingOrSelfPermission() to guard external caller methods. Always use enforceCallingPermission() or checkCallingPermission().
  • NEVER use deprecated signatureOrSystem.
  • NEVER declare or retain <grant-uri-permission android:pathPrefix="/" /> or pathPrefix="". Always replace with a scoped sub-path.
  • NEVER create infinite permission request loops on denial.
  • NEVER request foreground and background location permissions in the same request.
  • NEVER cache permission grant states in in-memory fields.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Android Permissions Security AI skill do?

Audits, detects gaps, and remediates Android permissions and IPC component security vulnerabilities. Use when reviewing AndroidManifest.xml, defining custom permissions, securing Services, Receivers, and Providers, verifying caller signatures and UID using Binder, or implementing runtime permission flows.

Why use Android Permissions Security on TypingMind?

Because you install it once and use it with any model. Android Permissions Security is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Android Permissions Security in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/android/skills/tree/main/security/android-permissions-security. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Android Permissions Security?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Android Permissions Security?

As many as you like. As long as a model supports skills, you can use Android Permissions Security with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Android Permissions Security AI skill free?

Yes. It is published on GitHub by android under the Apache-2.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇