The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The skill utilizes the dynamic command execution syntax to gather project metadata (engine type, git logs, file counts) during skill initialization. These commands are limited to reconnaissance of the current project directory and do not involve network operations, sensitive file access (e.g., .ssh or .env), or unsafe argument injection.
[INDIRECT_PROMPT_INJECTION]: The skill ingests file system metadata and git history into the agent's context to guide the onboarding workflow. While this data serves as an input surface, the skill implements it as a diagnostic function within a developer-focused context, and no exploitable capability chains were identified.

game-start