The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides instructions for the agent to modify Zsh configuration files, which presents an attack surface for indirect prompt injection if those files or user inputs contain malicious instructions.\n
Ingestion points: The agent is directed to read and edit .zshenv, .zprofile, .zshrc, .zlogin, and .zlogout (specified in SKILL.md and references/startup-files.md).\n
Boundary markers: The skill does not provide specific delimiters or warnings for the agent to ignore embedded instructions within these configuration files, though it does mention respecting {{ ... }} Chezmoi template blocks.\n
Capability inventory: The skill's primary function is to guide the agent in performing file-write operations to persistent shell configuration files across all referenced scripts.\n
Sanitization: There are no instructions for the agent to sanitize or validate the content of the configurations it is writing, such as user-provided aliases or environment variable values.

configuring-zsh