skills/simoneavogadro/android-reverse-engineering-skill/android-reverse-engineering/Gen Agent Trust Hub
android-reverse-engineering
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill includes scripts (
install-dep.sh,install-dep.ps1) that download reverse engineering tools includingjadx,vineflower, anddex2jardirectly from their official GitHub releases. - [COMMAND_EXECUTION]: The skill utilizes a variety of shell and PowerShell scripts to orchestrate the decompilation process, search for API endpoints, and recover Kotlin class names from metadata.
- [PRIVILEGE_ESCALATION]: The installation scripts may request elevated privileges (
sudo) to install required system packages like the Java JDK through standard package managers (apt, dnf, pacman). - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted Android package files (APK, XAPK, JAR). Data extracted from these files, such as string literals or decompiled code, is processed by the agent and could potentially contain malicious instructions intended to influence the agent's output or actions.
- Ingestion points:
scripts/decompile.shandscripts/fingerprint.shaccept path arguments to untrusted application files. - Boundary markers: The skill does not explicitly use boundary markers to separate untrusted data from the agent's instructions during the analysis phase.
- Capability inventory: The skill has access to the filesystem (read/write), network (for dependency updates), and command execution.
- Sanitization: No specific sanitization is applied to the content of the target applications before analysis.
Audit Metadata