Skills
skills/supercent-io/skills-template/firebase-cli/Gen Agent Trust Hub

firebase-cli

Fail

Audited by Gen Agent Trust Hub on Mar 13, 2026

Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches installation scripts from the well-known domain https://firebase.tools, which is an official resource provided by Google.
  • [REMOTE_CODE_EXECUTION]: The skill uses a piped bash execution pattern (curl -sL https://firebase.tools | bash) for installing the CLI. As this originates from a trusted organization, it is documented as intended behavior.
  • [COMMAND_EXECUTION]: Helper scripts scripts/deploy.sh and scripts/emulators.sh dynamically construct and execute commands using the eval function.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface due to the use of eval in shell scripts with variable interpolation of unsanitized user input. 1. Ingestion points: Command-line arguments passed to scripts/deploy.sh and scripts/emulators.sh (e.g., deployment messages). 2. Boundary markers: No delimiters or markers are present to isolate user data. 3. Capability inventory: Execution of firebase CLI commands via eval in the bash environment. 4. Sanitization: No sanitization or escaping of special characters is performed on user-provided strings before they are evaluated.
Recommendations
  • HIGH: Downloads and executes remote code from: https://firebase.tools - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Mar 13, 2026, 03:04 AM