The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted SQL migration files and schema definitions, creating a potential surface for indirect prompt injection.
Ingestion points: Migration files and schema paths provided via arguments in SKILL.md.
Boundary markers: No explicit delimiters or 'ignore' instructions are used to isolate processed SQL from the agent's instructions.
Capability inventory: Executes scripts/schema-diff.mjs and utilizes Bash and Task tools.
Sanitization: Includes a safety protocol requiring users to confirm destructive steps by typing the name of the affected object verbatim, and instructs the agent to redact secrets from findings.
[COMMAND_EXECUTION]: The skill invokes a local script scripts/schema-diff.mjs using Node.js to calculate schema differences. This is a functional requirement of the tool's intended use case and utilizes vendor-provided scripts.

migrate