The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches a live catalog of available tools from a remote DigitalOcean Spaces bucket at https://reeceyang.sgp1.cdn.digitaloceanspaces.com/SKILL.md. It also directs the agent to install the cli-anything-hub package from a public registry.
[REMOTE_CODE_EXECUTION]: The skill's primary function is to install and execute Python packages (cli-anything-hub and various cli-anything-<name> tools) on the local system using pip.
[COMMAND_EXECUTION]: The instructions rely on shell commands to manage software, including searching, listing, and installing packages, which grants the skill the ability to modify the environment and run arbitrary code associated with those packages.
[INDIRECT_PROMPT_INJECTION]: The skill encourages the agent to ingest and act upon data from an external catalog, which could contain instructions intended to influence the agent's behavior.
Ingestion points: The 'Live Catalog' section in SKILL.md directs the agent to read external tool descriptions and usage patterns from a remote URL.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the documentation for the catalog data.
Capability inventory: The skill utilizes shell command execution via pip install and cli-hub subcommands.
Sanitization: No validation or sanitization of the remote catalog content is specified before the agent processes it.

cli-hub-meta-skill