The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill facilitates the discovery and installation of external code packages from GitHub repositories using the npx skills add <package> command. This allows arbitrary third-party code to be executed within the agent's environment.
[COMMAND_EXECUTION]: The instructions direct the agent to execute shell commands (npx skills find, npx skills add, npx skills update) to manage capabilities. It specifically instructs the agent to use the -y flag, which bypasses user confirmation prompts, potentially leading to the silent installation of malicious software.
[EXTERNAL_DOWNLOADS]: The skill connects to the skills.sh registry and various GitHub repositories to fetch metadata and install package contents. While it references trusted sources like Vercel Labs, it also allows installation from any user-provided or search-discovered repository.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting data from external search results and user-provided queries.
Ingestion points: Search results from the skills.sh registry (via npx skills find) and user-supplied search queries.
Boundary markers: No delimiters or safety warnings are present when the agent interpolates user input into shell commands.
Capability inventory: Shell command execution via npx, network operations for discovery, and file system modification during installation.
Sanitization: The skill lacks validation or escaping for user-provided query strings before they are executed as part of shell commands.

find-skills