The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes several shell commands for environment verification and project management.
Runs docker info to check the status of the local Docker daemon.
Executes project-specific scripts (./gt build, ./gt run, ./gt gcp setup, ./gt start) to manage the application lifecycle.
Uses grep, cut, and tr to extract configuration values from local environment files.
[EXTERNAL_DOWNLOADS]: Triggers a container build process via ./gt build which typically involves fetching base images and dependencies from official package and image registries.
[DATA_EXPOSURE]: Interacts with local configuration files including scripts/gcp-config.env and checks for the existence of .gcloud/credentials.db. These actions are restricted to verifying setup state and do not involve exfiltrating data to external domains.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from scripts/gcp-config.env to determine the project ID. This is a low-risk ingestion point as the file is part of the local project environment and the extracted value is used for legitimate setup commands.
[SAFE]: The skill instructs the user to manually perform sensitive tasks like GCP authentication and OAuth client creation, ensuring that high-privilege operations remain under direct human control.

setup