The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill interacts with an external ecosystem where third-party skill descriptions could contain malicious instructions designed to influence the agent's behavior.
Ingestion points: Search results and metadata fetched via the npx skills find command described in SKILL.md.
Boundary markers: The skill includes a 'Verify Quality' step that instructs the agent to check install counts, GitHub stars, and source reputation.
Capability inventory: The skill can execute subprocesses via npx to install (add) or update packages.
Sanitization: The agent is instructed to present findings to the user and obtain explicit permission before performing any installation.
[COMMAND_EXECUTION]: Utilizes the Skills CLI to perform various lifecycle operations, including searching for packages and managing global installations.
[EXTERNAL_DOWNLOADS]: Downloads and installs agent extensions from external repositories. The instructions emphasize fetching content from official and well-known organizations.

find-skills