The Agent Skills Directory

[SAFE]: The skill primarily uses official tools and repositories from Google and Firebase to manage database instances and configurations.
[COMMAND_EXECUTION]: The skill uses npx -y firebase-tools@latest for managing Firestore databases, listing locations, and deploying rules and indexes. This is standard developer practice for Firebase projects.
[EXTERNAL_DOWNLOADS]: The skill fetches version metadata from the official Google Maven repository (dl.google.com) to ensure the agent uses deterministic, up-to-date SDK versions. These are trusted sources.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface where it instructs the agent to process local project files (such as build.gradle.kts and firebase.json) and extract documentation from local dependency caches (-sources.jar). However, these operations are scoped to project maintenance and documentation retrieval, and the instructions include validation steps like using gradlew to verify resolved versions.
[PROMPT_INJECTION]: The skill includes strong instructional constraints (e.g., "MANDATORY PIPELINE ARCHITECTURE", "ZERO TOLERANCE", "CRITICAL RULE") to enforce specific architectural patterns for Firestore Enterprise Native mode. These are functional requirements for the vendor's platform and do not attempt to override the agent's safety or ethical guidelines.

firebase-firestore