skills/mukul975/anthropic-cybersecurity-skills/exploiting-deeplink-vulnerabilities/Gen Agent Trust Hub
exploiting-deeplink-vulnerabilities
Pass
Audited by Gen Agent Trust Hub on Mar 15, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill uses
adb(Android Debug Bridge) andapktoolto interact with mobile devices and decompile APK files. This behavior is expected and necessary for its primary purpose of mobile security testing. - [INDIRECT_PROMPT_INJECTION]: The
scripts/process.pyandscripts/agent.pyfiles parse external data sources such asAndroidManifest.xmlandInfo.plist. These files are treated as untrusted input that influences the generation of testing commands. While this creates a theoretical attack surface for indirect injection, it is a standard characteristic of security analysis tools and does not pose a direct threat in this context. - [EXTERNAL_DOWNLOADS]: The skill documentation mentions downloading
apple-app-site-associationfiles viacurlfrom target domains for verification purposes. This is a legitimate part of the universal link assessment workflow.
Audit Metadata