The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill provides instructions for installing and running external code using the npx skills add command. This capability allows the agent to fetch and execute packages from arbitrary GitHub repositories or other remote sources.
[COMMAND_EXECUTION]: The skill uses the npx tool to execute shell commands. It specifically encourages the use of the -y flag (npx skills add <package> -g -y) to skip user confirmation during the installation process, which reduces the opportunity for human oversight of executed code.
[EXTERNAL_DOWNLOADS]: The skill directs the agent to download and interact with content from https://skills.sh/ and various external GitHub repositories. While it references well-known organizations, the tool is designed to install skills from any source, posing a potential supply-chain risk.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because it processes untrusted data from external search results. 1. Ingestion points: External search results returned by npx skills find and metadata from https://skills.sh/ (SKILL.md). 2. Boundary markers: Absent. 3. Capability inventory: Subprocess execution of shell commands to install packages (SKILL.md). 4. Sanitization: The skill includes a 'Verify Quality Before Recommending' step which instructs the agent to check install counts and source reputation (SKILL.md).

find-skills