The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches additional agent skills and the Firebase CLI (firebase-tools) from official sources via npx. These are well-known resources managed by the Firebase and Google organizations.
[COMMAND_EXECUTION]: The skill requires the execution of multiple shell commands to verify the environment, perform user authentication, and manage project lifecycles.
[DATA_EXFILTRATION]: Directs the agent to read and modify sensitive local configuration files (e.g., claude_desktop_config.json, mcp_config.json) to register the Firebase MCP server. This is a functional requirement for the skill but involves accessing protected configuration paths.
[PROMPT_INJECTION]: The skill provides a surface for indirect prompt injection due to the lack of input handling for user data.
Ingestion points: User-provided <PROJECT_ID> and <DISPLAY_NAME> values in SKILL.md.
Boundary markers: Absent; there are no delimiters or instructions to the agent to treat these inputs as potentially untrusted data.
Capability inventory: The skill possesses command execution capabilities through npx and the Firebase CLI.
Sanitization: User inputs are interpolated directly into shell command arguments without evidence of sanitization, validation, or escaping.

firebase-basics