rye-agent-ops

What problem does it solve?

LLM agents interacting with Rye-managed PostgreSQL data often violate row-level security policies, assertion lifecycle rules, and audit requirements, leading to data leaks, corrupted knowledge states, and non-compliant operations.

Core Features & Use Cases

• Safe Read/Write Guidance: Directs agents to use official Rye SQL helper functions instead of raw queries, enforcing RLS, team boundaries, and immutable assertion rules.
• Auditable Workflows: Covers event logging, provenance tracking, and knowledge candidate promotion flows to ensure all agent actions are traceable and compliant.
• Domain Integration: Helps agents connect external domain tables to the Rye graph and track table changes via CDC triggers. For example, a sales agent can safely ingest CRM data into Rye, create validated knowledge candidates, and schedule future deal stage changes without bypassing access controls.

Quick Start

Use the rye-agent-ops skill to set up your Rye session context, retrieve scoped node information, and safely record a new event with participant details for your workflow.