The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions direct the agent to execute shell commands (npx skills find and npx skills add) based on user-provided queries and package names.
[REMOTE_CODE_EXECUTION]: The npx skills add <package> command downloads and installs external 'skills' from remote GitHub repositories. This allows for the execution of untrusted third-party code on the host system. The use of the -y flag (skip confirmation) and -g flag (global installation) increases the risk of automatic execution of malicious code if a user is social-engineered into installing a specific package.
[EXTERNAL_DOWNLOADS]: The skill facilitates the retrieval of external packages, templates, and workflows from GitHub and the skills.sh registry.
[COMMAND_EXECUTION]: User input is interpolated directly into shell commands (npx skills find [query]). The absence of explicit sanitization or shell escaping for the [query] parameter introduces a command injection vulnerability where a user could supply shell metacharacters (e.g., ;, &&, |) to execute arbitrary commands.
[PROMPT_INJECTION]: The skill processes untrusted data from external sources (skill names and descriptions returned by the search command). This represents an indirect prompt injection surface where a malicious skill description could influence the agent's subsequent actions.

find-skills