generated: '2026-07-21'
method: searched
source: https://github.com/RelationalAI/rai-agent-skills
notes: >-
  RelationalAI PUBLISHES its own set of Agent Skills (Apache-2.0) in the
  rai-agent-skills repository — markdown skills coding agents load at runtime to
  drive the decision-intelligence workflow against the `relationalai` (PyRel)
  Python SDK. Installable into 45+ agents (Claude Code, Cursor, Copilot, Codex,
  Cortex Code). Prerequisites: relationalai v1.20.1+, the RelationalAI Native
  App installed in Snowflake, and the rai_developer role. These are the
  provider's own published skills, referenced verbatim by source URL rather than
  regenerated.
repo: https://github.com/RelationalAI/rai-agent-skills
api: python:relationalai
skills:
- file: skills/rai-setup
  name: rai-setup
  area: setup
  purpose: Configuration, Snowflake connection, and engine management.
  source: https://github.com/RelationalAI/rai-agent-skills/tree/main/skills/rai-setup
- file: skills/rai-pyrel
  name: rai-pyrel
  area: development
  purpose: PyRel v1 syntax, data loading, and query construction.
  source: https://github.com/RelationalAI/rai-agent-skills/tree/main/skills/rai-pyrel
- file: skills/rai-ontology
  name: rai-ontology
  area: ontology
  purpose: Build and evolve domain models from data.
  source: https://github.com/RelationalAI/rai-agent-skills/tree/main/skills/rai-ontology
- file: skills/rai-discovery
  name: rai-discovery
  area: reasoning
  purpose: Surface answerable questions and classify by reasoner type.
  source: https://github.com/RelationalAI/rai-agent-skills/tree/main/skills/rai-discovery
- file: skills/rai-graph-analysis
  name: rai-graph-analysis
  area: reasoning
  purpose: Centrality, community detection, and reachability algorithms.
  source: https://github.com/RelationalAI/rai-agent-skills/tree/main/skills/rai-graph-analysis
- file: skills/rai-predictive-modeling
  name: rai-predictive-modeling
  area: reasoning
  purpose: Graph neural network (GNN) model construction.
  source: https://github.com/RelationalAI/rai-agent-skills/tree/main/skills/rai-predictive-modeling
- file: skills/rai-predictive-training
  name: rai-predictive-training
  area: reasoning
  purpose: Train GNNs, generate predictions, and evaluate results.
  source: https://github.com/RelationalAI/rai-agent-skills/tree/main/skills/rai-predictive-training
- file: skills/rai-prescriptive-problem
  name: rai-prescriptive-problem
  area: reasoning
  purpose: Formulate optimization and constraint-satisfaction (CSAT) problems.
  source: https://github.com/RelationalAI/rai-agent-skills/tree/main/skills/rai-prescriptive-problem
- file: skills/rai-prescriptive-results
  name: rai-prescriptive-results
  area: reasoning
  purpose: Execute solves, interpret output, and run sensitivity analysis.
  source: https://github.com/RelationalAI/rai-agent-skills/tree/main/skills/rai-prescriptive-results
- file: skills/rai-deployment
  name: rai-deployment
  area: operations
  purpose: Production deployment to Snowflake and lifecycle management.
  source: https://github.com/RelationalAI/rai-agent-skills/tree/main/skills/rai-deployment
- file: skills/rai-health
  name: rai-health
  area: operations
  purpose: Diagnose engine performance and data-stream health.
  source: https://github.com/RelationalAI/rai-agent-skills/tree/main/skills/rai-health