The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill provides commands to install the Valet CLI by downloading and executing shell scripts directly from the vendor's website (valet.dev) via curl -fsSL https://valet.dev/install.sh | sh and irm https://valet.dev/install.ps1 | iex. While this is the official installation path for the vendor, it involves piping remote code to a shell.\n- [COMMAND_EXECUTION]: The skill makes extensive use of the terminal to run valet CLI commands for agent lifecycle management and uses valet exec to execute local commands with secrets injected into the process environment.\n- [INDIRECT_PROMPT_INJECTION]: The skill supports the creation of agents that ingest untrusted data via webhooks. It provides guidance to include boundary markers and strict parsing instructions to reduce the risk of the agent obeying instructions embedded in external payloads.\n
Ingestion points: Webhook JSON payloads appended to channel instruction files.\n
Boundary markers: Present (instructions for the agent to parse payloads inline and ignore external tool calls for finding data).\n
Capability inventory: Shell command execution via valet exec, agent deployment, and secret management.\n
Sanitization: Absent.\n- [EXTERNAL_DOWNLOADS]: The skill fetches installation scripts from valet.dev and instructs the agent to use WebFetch to discover external tool configurations and documentation from github.com and skills.sh.

valet