The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute shell commands using the npx skills CLI tool for searching, checking, and updating installed packages.
[EXTERNAL_DOWNLOADS]: The skill enables the download and installation of functional modules from external sources such as GitHub and the skills.sh registry.
[REMOTE_CODE_EXECUTION]: The core functionality involves installing and potentially running third-party agent skills, which constitutes remote code execution. The documentation suggests using the -y flag to skip interactive confirmation prompts when the user expresses intent to install a skill.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes data from an external ecosystem (package descriptions and search results) that could be controlled by an attacker.
Ingestion points: Output from the npx skills find command processed at runtime.
Boundary markers: None identified; the agent is instructed to present discovered skill descriptions directly to the user.
Capability inventory: Subprocess execution and code installation via npx skills add.
Sanitization: No validation or sanitization of external metadata (e.g., skill names or descriptions) is specified before the agent acts on the data.

find-skills