Context Engineering Studio Private Preview
A Context Repo is a versioned, domain-scoped unit of context any AI agent can consume. It bundles skills (how the agent should reason), a semantic model (what your metrics mean), and deploy and eval configs into one governed package.
- Generated from your data graph. Context Engineering Studio reads your catalog, BI lineage, and query history to generate metrics, filters, and relationships automatically.
- Evaluated before deployment. Run evals against the repository to surface gaps with specific fix recommendations before it goes live.
- Deploy to any agent or engine. One certified repository serves MCP-compatible agents, Snowflake Cortex, and Databricks Genie, with live query traces for continuous iteration.
context-repo
v3.1.2 · domain repository · MCP-ready
Contents
semantic-model.yaml
skills
finance-compliance.md
analyst.md
deploy
databricks.json
quality-report.md
Finance Compliance
Role & context
Guides the agent when a question touches regulated or sensitive data — enforcing approved sources, masking, and disclosure rules before any answer is returned.
Guardrails
- Only use certified, governed tables as sources
- Never expose PII or restricted columns in answers
- Cite the source asset for every figure returned
Applies to
Restricted tagsPII, Confidential
Scopefinance domain
2 skills1 semantic model1 deploy targetquality 82%
1
Enable
Enable CES on your tenant, assign access to your team, and connect your query engine.
→
2
Build
Describe your domain, generate a semantic model, refine definitions, and simulate with a question set to see where to improve.
→
3
Deploy
Certify your context repository and deploy it to Snowflake Cortex Analyst or Databricks Genie.