The Agent Skills Directory

[Remote Code Execution] (CRITICAL): The file SKILL.md contains a command to download and execute a script directly from an external URL: curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.40.3/install.sh | bash. This piped execution pattern is highly dangerous as it grants the remote script full execution rights without prior inspection. The source nvm-sh is not on the list of trusted GitHub organizations.
[Command Execution] (HIGH): The SKILL.md file suggests using sudo to resolve EACCES errors during global npm installations. This practice promotes privilege escalation and unnecessarily increases the attack surface of the system.
[Indirect Prompt Injection] (LOW): The skill creates a surface for indirect prompt injection by facilitating the ingestion of user-controlled strings (Project IDs, app names) into CLI tool parameters without sanitization or boundary markers.
Ingestion points: Project ID and display name prompts in firebase projects:create and firebase apps:create in SKILL.md and references/web_setup.md.
Boundary markers: Absent; user inputs are interpolated directly into commands.
Capability inventory: CLI execution for project management, authentication, and directory creation.
Sanitization: None present.

firebase-basics