The Agent Skills Directory

[SAFE]: The skill is primarily instructional, offering architectural patterns for Kotlin Multiplatform development. It focuses on code organization and abstraction decisions for the Amethyst project.
[COMMAND_EXECUTION]: The skill contains local helper scripts (scripts/suggest-kmp-dependency.sh and scripts/validate-kmp-structure.sh). These scripts utilize standard Unix utilities such as grep, find, sed, and awk to perform static analysis of the local project's source code and Gradle configurations. They do not initiate network connections, download external content, or modify sensitive system files.
[EXTERNAL_DOWNLOADS]: No remote script downloads or executions (e.g., curl | bash) are present. Dependency references involve standard, well-known libraries (Jackson, OkHttp, Ktor, Kotlinx) typically managed via the project's own version catalog.
[PROMPT_INJECTION]: The skill does not contain instructions that attempt to override the AI agent's safety protocols or extract system prompts.
[DATA_EXFILTRATION]: No evidence was found of attempts to access sensitive user files, credentials, or environment variables for the purpose of exfiltration.

kotlin-multiplatform