The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is centered around executing the pscale CLI to manage cloud database resources. This includes operations with significant impact, such as pscale pgbouncer delete (which includes a --force flag) and pscale pgbouncer resize. The skill mitigates this by requiring user approval for destructive or capacity-changing operations.
[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and act upon data returned from the pscale CLI in JSON format. This creates a surface for indirect prompt injection if the external API returns malicious content.
Ingestion points: Commands such as pscale pgbouncer list, pscale pgbouncer show, and pscale pgbouncer resize status return data from the PlanetScale API that is processed by the agent.
Boundary markers: The instructions do not define explicit boundary markers or warnings to the agent to treat the CLI output as untrusted data.
Capability inventory: The skill has the capability to modify infrastructure (resize, create, delete) via the pscale CLI.
Sanitization: There are no explicit sanitization or validation steps defined for the data returned by the CLI before it is used to inform subsequent actions.

pscale-pgbouncer