android-pentesting-tricks
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes several command-line utilities for mobile security research, including Android Debug Bridge (adb), aapt, and the Frida/Objection instrumentation frameworks. These tools are used to interact with Android devices, push files, launch activities, and inspect application manifests.
- [DYNAMIC_EXECUTION]: The skill includes a collection of Frida script templates (in FRIDA_SCRIPTS.md) designed to perform runtime instrumentation. These scripts hook Java methods within the Android runtime to bypass security controls like SSL pinning and root detection, or to extract encryption keys and application data.
- [INDIRECT_PROMPT_INJECTION]: As the skill is designed to ingest and analyze untrusted data from Android applications (such as manifest files, intent data, and database content), it possesses an inherent attack surface for indirect prompt injection. A malicious application could provide crafted inputs to influence the behavior of the agent during the analysis process.
- Ingestion points: ADB backup data, AndroidManifest.xml files, Content Provider query results, and Intent extras (SKILL.md).
- Boundary markers: None present.
- Capability inventory: File system interaction via adb, instrumentation via Frida, and command execution (SKILL.md, FRIDA_SCRIPTS.md).
- Sanitization: None present.
Audit Metadata