The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill facilitates the installation of code packages from arbitrary GitHub repositories using the npx skills add command. This allows the download and execution of remote code from sources that may not be verified.
[COMMAND_EXECUTION]: Utilizes the npx utility to execute the skills CLI tool, passing user-supplied search queries and repository names as command-line arguments.
[INDIRECT_PROMPT_INJECTION]: Takes natural language input from users to construct search queries for the skills CLI. This provides an attack surface for indirect prompt injection if malicious data is provided in a request for help.
Ingestion points: User requests for task assistance or skill discovery.
Boundary markers: None present; user input is directly interpolated into the command string.
Capability inventory: Includes shell command execution and writing files to the local llmwiki/wiki/ directory.
Sanitization: The instructions do not describe any input validation or sanitization steps before the query is processed by the CLI tool.
[EXTERNAL_DOWNLOADS]: Mentions downloading configuration and documentation from skills.sh and specific repositories from Vercel Labs and Anthropic. These references are part of the skill's package management functionality.

find-skills