The Agent Skills Directory

[SAFE]: The skill documents architectural best practices for maintaining parity between staging and production environments while utilizing ephemeral resources for development and CI.
[EXTERNAL_DOWNLOADS]: Utilizes well-known official Docker images (e.g., postgres, redis, mailhog, minio) from public registries for infrastructure services.
[COMMAND_EXECUTION]: Employs standard development commands for environment setup and lifecycle management, such as docker compose, npm run, and npx for test runners.
[PROMPT_INJECTION]: Ingests project-specific constraints from a local context file (.agents/qa-project-context.md), which is a standard surface for aligning agent behavior with project requirements.
[CREDENTIALS_UNSAFE]: Uses mock/test-mode credentials (e.g., sk_test_fake) and default container passwords which are restricted to isolated test environments.

test-environments