live wire
▸Oracle Select AI generates SQL/PGQ for property graphs—but errors if one AI profile mixes graphs with tables or viewsOracle▸Domo’s data team standardizes MCP builds with shared scaffolding; beta toolkits inherit user permissionsDomo▸Databricks proposes five outcome signals for agentic BI—and says dashboard counts and license usage miss decision qualityDatabricks▸One vibe-coded database app needed 56+ human test hours and three repair rounds despite AI Studio’s verification claimsMancas & Mancas / arXiv▸BigQuery AI.IF’s default cost mode trains a local model for most rows—and drops row-error thresholdsGoogle Cloud▸AWS’s multi-tenant data-agent pattern gives the broker Lambda no data permissions; Cognito roles carry Lake Formation scopeAWS Big Data Blog▸Amazon OpenSearch’s optimized engine adds nested agent traces, but AWS reserves it for log analytics—not vector searchAWS▸Two interpretation-based provenance semantics cannot express SQL bag difference, new theorem showsKolaitis / arXiv▸Databricks Feature Views reuse one definition for training and live scoring, but online lookup still returns latest available stateDatabricks▸Databricks Genie’s Slack and Teams previews default channel answers to public, even when viewers lack source-data accessDatabricks▸Snowflake’s two-tool GraphRAG agent beat a seven-tool graph setup 70% to 60% on 22 ontology questionsSnowflake Engineering▸Semantix opens an Apache-2.0 semantic-first data-agent stack, but labels the project WIP and has no release or adoption yetGitHub / icosmos-space▸Confluent defines Streamhouse as shared current state for agents and apps, but says broader semantics and governance are still to comeConfluent▸Graph-repair retrieval cut F1 by 0.126 on one suite; a learned router used fewer tokens than always-on retrieval in all sixLin et al. / arXiv▸Oracle Select AI generates SQL/PGQ for property graphs—but errors if one AI profile mixes graphs with tables or viewsOracle▸Domo’s data team standardizes MCP builds with shared scaffolding; beta toolkits inherit user permissionsDomo▸Databricks proposes five outcome signals for agentic BI—and says dashboard counts and license usage miss decision qualityDatabricks▸One vibe-coded database app needed 56+ human test hours and three repair rounds despite AI Studio’s verification claimsMancas & Mancas / arXiv▸BigQuery AI.IF’s default cost mode trains a local model for most rows—and drops row-error thresholdsGoogle Cloud▸AWS’s multi-tenant data-agent pattern gives the broker Lambda no data permissions; Cognito roles carry Lake Formation scopeAWS Big Data Blog▸Amazon OpenSearch’s optimized engine adds nested agent traces, but AWS reserves it for log analytics—not vector searchAWS▸Two interpretation-based provenance semantics cannot express SQL bag difference, new theorem showsKolaitis / arXiv▸Databricks Feature Views reuse one definition for training and live scoring, but online lookup still returns latest available stateDatabricks▸Databricks Genie’s Slack and Teams previews default channel answers to public, even when viewers lack source-data accessDatabricks▸Snowflake’s two-tool GraphRAG agent beat a seven-tool graph setup 70% to 60% on 22 ontology questionsSnowflake Engineering▸Semantix opens an Apache-2.0 semantic-first data-agent stack, but labels the project WIP and has no release or adoption yetGitHub / icosmos-space▸Confluent defines Streamhouse as shared current state for agents and apps, but says broader semantics and governance are still to comeConfluent▸Graph-repair retrieval cut F1 by 0.126 on one suite; a learned router used fewer tokens than always-on retrieval in all sixLin et al. / arXiv
nl2sql.ai
guideOperations guide

AWS put agent discovery behind an approval gate—but that does not authorize the tool at runtime

Agent Registry is now in GovCloud US-West. The useful boundary is also the easy one to miss: catalog approval governs discoverability, while execution still belongs to separate runtime, gateway and identity controls.

By The News Desk· Nov 6, 2026the quick take — two AI hosts go live when you do

AWS has made Agent Registry generally available in GovCloud US-West, giving public-sector teams a managed catalog for MCP servers, tools, agents, skills and custom resources. The launch adds an approval workflow, a native MCP endpoint and CloudTrail audit logging to the region (AWS launch).

The important operational point is what the registry does not do. Approval decides which records become discoverable; it is not a blanket authorization for an agent to execute the underlying tool.

Treat the catalog and the runtime as separate control planes

AWS describes a workflow in which publishers submit records, curators approve or reject them, and consumers search or browse only the approved catalog. Administrators can also deprecate records that should no longer be found (Agent Registry documentation). That is a strong control for inventory quality: builders and agents can find an approved resource instead of relying on an unreviewed endpoint copied into a prompt or configuration file.

But a registry record is metadata about a resource. AWS documents separate services for hosting agents and MCP servers, converting APIs into MCP-compatible tools, and managing identity and credentials. The registry itself can authenticate search and MCP access with IAM credentials or corporate-identity JWTs (Agent Registry documentation). Those boundaries mean a discoverable tool still needs its own invocation permissions and data-access policy.

For a data agent, this distinction matters. A registry approval can establish that an MCP server is the organization’s sanctioned interface for analytics, but it does not prove that every user or agent should be able to query every database behind it. Keep warehouse roles, row- and column-level controls, and write permissions attached to the runtime identity rather than treating catalog membership as the grant.

A practical deployment checklist

Use the registry approval as the publication gate: verify the owner, endpoint, schema, intended environment and support status before a record becomes discoverable. Separate development, QA and production registries—or at least encode the lifecycle stage clearly—because AWS allows registries to be organized by environment, team or resource type (Agent Registry documentation).

Then test the second gate independently. Confirm that an approved record can be discovered by the intended principals, while the underlying tool rejects unauthorized invocation. Log registry API activity in CloudTrail, but retain runtime and downstream data-system logs as well; the regional launch specifically identifies CloudTrail as the registry audit surface (AWS launch).

The result is a cleaner model: approval says this is the tool we endorse; runtime authorization says this principal may use it for this action. GovCloud availability makes the first control easier for regulated teams, but it does not replace the second.

Filed by The News Desk. Corrections: desk@nl2sql.ai · Our standards →

comments · 0

    Comments are moderated before they appear. Your email is used once to confirm it is you — never shown, never sold. Corrections and questions get an answer from the desk when we have one.