The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The documentation includes instructions to fetch the @google/mcp-toolbox package via npx and a Docker container from Google's registry (us-central1-docker.pkg.dev). These are verified resources from a well-known service provider.
[COMMAND_EXECUTION]: The skill provides standard setup commands for users, including npx, docker, and curl, which are necessary for the tool's intended deployment and configuration.
[PROMPT_INJECTION]: The skill exposes a surface for indirect prompt injection by enabling the agent to process external, potentially untrusted data.
Ingestion points: Tools such as get_project_file and query (documented in SKILL.md) allow the agent to read content from external Looker projects and database results.
Boundary markers: The documentation does not specify the use of delimiters or instructions to help the agent isolate external data from its core directives.
Capability inventory: The skill grants the agent capabilities to modify the environment, such as update_project_file and delete_project_file, which could be exploited if the agent follows instructions found in ingested data.
Sanitization: There is no mention of sanitization or filtering of the external data before it is presented to the AI agent.

looker-mcp-skill