The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides documentation for various CLI commands including database schema management and raw SQL execution (prisma db execute). It handles destructive actions such as prisma migrate reset and prisma db push --force-reset with specific warnings.
[DYNAMIC_EXECUTION]: Documents standard ORM workflows that involve code generation (prisma generate) and the execution of user-defined database seeding scripts using runtime tools like tsx, ts-node, or node.
[INDIRECT_PROMPT_INJECTION]: As a technical reference, the skill exposes the agent to data ingestion surfaces such as database schema files (.prisma) and SQL scripts. The risk is mitigated by explicit instructions requiring the agent to obtain user consent and follow an 'AI safety checkpoint' protocol before taking sensitive actions.
[SAFE]: The skill implements a robust safety mechanism (references/agent-safety.md) designed to prevent agents from performing destructive operations without verified user intent, using a specific environment variable (PRISMA_USER_CONSENT_FOR_DANGEROUS_AI_ACTION) to track consent.

prisma-cli