The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes various local scripts and platform-specific CLI tools to perform builds and deployments.
Evidence: Invokes project-specific scripts including ./scripts/macos_appstore_upload.sh, ./scripts/ios_appstore_upload.sh, ./scripts/android_playstore_upload.sh, and ./scripts/deploy_all.sh.
Evidence: Executes standard development tools such as fastlane, flutter, fvm, and firebase.
[DATA_EXFILTRATION]: The skill accesses sensitive credential files required for app store authentication.
Evidence: Reads Apple API keys from the ~/.app_store_credentials/AuthKey_*.p8 directory and accesses environment variables like $APP_STORE_CONNECT_KEY_ID.
Evidence: Accesses local provisioning profiles in ~/Library/Developer/Xcode/UserData/Provisioning Profiles/.
Evidence: Reads project configuration secrets from .env files using the --dart-define-from-file flag.
[PROMPT_INJECTION]: The skill processes external project files which presents a surface for indirect prompt injection.
Ingestion points: Reads content from pubspec.yaml, .arb localization files, and description.txt metadata files (SKILL.md).
Boundary markers: Absent; there are no instructions to ignore malicious directives that might be embedded in the ingested text.
Capability inventory: Includes full shell access for builds, git commit/push capabilities, and network upload capabilities to App Store and Google Play.
Sanitization: Absent; metadata content is read and used directly for store listing preparation without filtering.

release-app