The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill uses the Bash tool to execute npx agentshq. This command downloads and executes the agentshq package from the NPM registry at runtime.
[EXTERNAL_DOWNLOADS]: The add command (npx agentshq add <url>) allows the user to fetch agent definitions from remote repositories and URLs, introducing external content into the environment.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it installs agent definitions from external, potentially untrusted sources without sanitization.
Ingestion points: Agent repositories and URLs provided to the add command.
Boundary markers: None implemented for the content of downloaded agents.
Capability inventory: The skill has access to Bash, Read, and AskUserQuestion tools.
Sanitization: No validation or sanitization of the downloaded agent instructions is performed.
[COMMAND_EXECUTION]: The skill executes shell commands via npx to manage the lifecycle of AI agents across various IDEs and CLIs.

find-agents