The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes data from external sources that could contain malicious instructions.
Ingestion points: Data returned from PlanetScale CLI commands such as pscale database list, pscale database show, and pscale keyspace show as described in SKILL.md.
Boundary markers: The instructions do not define explicit boundary markers or delimiters for data ingested from CLI outputs to prevent the agent from following instructions embedded in the data.
Capability inventory: The skill possesses extensive capabilities including creating, updating, and deleting databases and keyspaces, and opening interactive shells (pscale database create, pscale database delete, pscale database update, pscale shell).
Sanitization: No explicit sanitization or validation of the CLI output is described, although the documentation recommends using --format json and parsing results with jq.
[COMMAND_EXECUTION]: The skill interacts directly with the pscale CLI tool to perform infrastructure management tasks.
The skill provides instructions for executing administrative commands such as pscale database delete, pscale keyspace delete, and pscale shell, which grant significant control over database environments.
The workflow for creating external keyspaces involves executing shell scripts that use environment variables for credential management. While the skill provides best-practice advice to avoid persisting secrets in files or chat history, these operations involve sensitive data handling and execution of multi-line shell commands.

pscale-database