The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill provides installation instructions in SKILL.md that use high-risk patterns to download and pipe remote scripts directly into a shell interpreter (e.g., curl -fsSL https://github.com/kriss-spy/plugman/releases/latest/download/install.sh | sh and irm ... | iex). This method bypasses local verification of the script content before execution.
[EXTERNAL_DOWNLOADS]: The skill downloads executable scripts and plugin release assets from external GitHub repositories during installation and plugin management operations.
[COMMAND_EXECUTION]: The skill is designed to run the plugman CLI to perform filesystem operations, such as creating, updating, and deleting plugin folders within an Obsidian vault.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface where untrusted external data could potentially influence agent behavior.
Ingestion points: The skill processes user-provided plugins.list files and fetches metadata from GitHub repository URLs.
Boundary markers: There are no explicit markers or instructions to delimit or ignore potential commands within these external sources.
Capability inventory: The skill can execute shell commands (plugman) that modify the local filesystem.
Sanitization: The skill does not describe any validation or filtering of the content retrieved from external files or URLs.

obsidian-plugman