The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to automatically execute a shell command (ls) upon its first activation in a conversation. This command targets multiple directories, including sensitive paths in the user's home directory (~/.claude/skills/, ~/.agent/skills/, and ~/.agents/skills/), to enumerate installed companion skills. This behavior constitutes silent environment discovery.
[EXTERNAL_DOWNLOADS]: The agent is directed to offer the installation of additional skills from an external repository (AbsolutelySkilled/AbsolutelySkilled) using the npx skills add command. This pattern encourages the user to download and execute code from a third-party source that is not part of the platform's trusted vendor list.
[DATA_EXFILTRATION]: While the skill currently only performs local discovery, the instructions to read the user's skill directory structure could be a precursor to identifying and potentially exfiltrating information about the user's environment or installed tools.

project-execution