The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of shell commands to automate the analysis workflow, utilizing utilities such as curl, grep, unzip, 7z, openssl, adb, and python3 for data processing and device interaction.
[EXTERNAL_DOWNLOADS]: The skill facilitates the acquisition of Android application packages (APKs) from third-party mirror services including APKPure and APKMirror. It also provides instructions for downloading the jadx decompilation tool from its official repository on GitHub. The use of the -k or --insecure flag with curl in multiple stages indicates that SSL certificate validation is bypassed during these downloads.
[CREDENTIALS_UNSAFE]: The skill includes a large library of regular expression patterns specifically designed to identify hardcoded sensitive information within application binaries, such as AWS Access Keys, Google API keys, GitHub tokens, and OpenAI API keys. While these are intended for detection during a red-team engagement, they represent a significant focus on credential discovery.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: The skill ingests and processes untrusted external data in the form of HTML content from the Play Store and compiled APK files from third-party mirrors.
Boundary markers: There are no explicit boundary markers or instructions to the agent to ignore embedded commands within the processed data.
Capability inventory: The skill possesses the capability to execute shell commands, perform network requests via curl, and interact with Android devices via adb and frida.
Sanitization: The skill performs direct string extraction and pattern matching using strings and grep on external binaries without intermediate sanitization or validation steps.

apk-redteam-pipeline