The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides several shell command examples designed to inspect the local project environment and build artifacts. These commands utilize standard utilities like find, grep, sed, awk, unzip, and diff to analyze the Gradle cache (~/.gradle/caches), project build scripts (build.gradle.kts), and Kotlin source files. The commands are intended for manual or agent-led diagnostic purposes within a development workflow.
[INDIRECT_PROMPT_INJECTION]: The skill establishes an attack surface by instructing the agent to process data from external files within the project directory, such as source code and build configuration files. Malicious content within these files could potentially influence the agent's logic when executing the diagnostic steps.
Ingestion points: Project files including build.gradle.kts, *.kt source files, and compiled .aar or .jar dependencies.
Boundary markers: No specific delimiters or safety warnings are provided for the output of the inspection commands.
Capability inventory: The skill utilizes file system traversal, text searching, and archive extraction capabilities.
Sanitization: The provided commands do not include specific sanitization or validation of the contents of the files being processed.

noop-actual-not-platform-limit