The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests data from project-specific files such as .meu/config.json, presets.dist.js, and presets.entry.js to determine code generation parameters. While this creates a surface for indirect prompt injection from potentially attacker-controlled local files, the skill's primary purpose is contextual code generation and it explicitly forbids script execution, mitigating the risk.\n- Ingestion points: Local configuration and preset files (e.g., .meu/config.json, presets.dist.js) are read from the project directory to resolve paths and exported signatures.\n- Boundary markers: The instructions do not define specific delimiters for separating user-provided instructions from the content of the project files being processed.\n- Capability inventory: The skill generates scripts for Android instrumentation, UI interaction, and state monitoring using Frida APIs.\n- Sanitization: The agent is not instructed to sanitize or validate the content of the project files before using them in the code generation process.

to-android-script