The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data that could potentially influence agent behavior if the underlying scripts do not perform strict validation. * Ingestion points: The device, module_path, and wait_text parameters in the magisk-setup.toon and install-module.toon workflows in SKILL.md. * Boundary markers: The workflow instructions lack explicit delimiters or safety prompts to prevent the agent from acting on instructions embedded within these parameters. * Capability inventory: The skill executes Python scripts (adb-magisk-launch.py, adb-magisk-enable-zygisk.py, adb-magisk-install-module.py) that perform file selection, UI interaction, and system configuration on the target Android device via ADB. * Sanitization: No input sanitization or validation logic is described in the skill documentation.
[EXTERNAL_DOWNLOADS]: The installation command fetches the skill definition from a public GitHub repository. * Evidence: metadata.json includes a curl command targeting raw.githubusercontent.com.

adb-magisk