The Agent Skills Directory

[COMMAND_EXECUTION]: The skill enables the agent to manage Render cloud services by executing CLI commands for deployment, workspace configuration, and log inspection.
[DATA_EXFILTRATION]: The skill includes instructions for using render psql and render ssh, which provide direct access to remote database contents and service environments.
[PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection as it retrieves and processes data from external sources such as the Render API and remote databases. * Ingestion points: The agent reads metadata from render services and query results from render psql. * Boundary markers: No delimiters are specified to distinguish between CLI output and agent instructions. * Capability inventory: The skill can trigger infrastructure changes via render deploys and access sensitive environments via render ssh. * Sanitization: There is no evidence of validation or filtering for content retrieved from external CLI outputs.

render