The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes the skilllint utility via the uvx command in a Bash environment to perform quality checks on skill directories.
[EXTERNAL_DOWNLOADS]: The skill triggers the download and execution of the skilllint package from a remote registry at runtime using the uvx tool.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from user arguments and external tool outputs, which constitutes an indirect prompt injection surface.
Ingestion points: User-provided <path> argument and the output generated by the uvx skilllint command.
Boundary markers: Not explicitly defined for the external tool output.
Capability inventory: Execution of shell commands through Bash and uvx.
Sanitization: The skill contains proactive security instructions that explicitly warn the agent not to splice user-supplied text into shell-interpreted strings or dynamic context injection patterns (e.g., !...) to prevent command injection.

lint