OpenLink Software

Here's our ODBC Driver Administrator Skill for creating and testing ODBC Data Source Names (DSNs). It's usable by any AI Agent and platform that supports the SKILLS.md open standard. Benefit? It makes ODBC easier to use, which means DBMS independence just got a lot simpler!

A very basic example of what attribute-based access control (#ABAC) looks like when using the OpenLink Request Broker to handle a JDBC connection requested from an ODBC client. Breakdown: The Request Broker receives a connection request for a JDBC 1.4 data source named LocalVirtuoso, using the user dba, originating from the isql application on localhost, with read/write access. It attempts to find a specific mapping for this request but falls back to a generic JDBC 1.4 mapping. Based on this mapping, it selects the generic_jdbc14 configuration, which specifies the Database Agent jdbc18_mv. Connection parameters are then prepared: Domain: JDBC 1.4 Database: LocalVirtuoso Read-only mode force: disabled (read/write) User: dba Operating system: macos Machine: localhost Application: isql Finally, the Request Broker spawns the JDBC Bridge Agent process:

Our multi-tier database connectivity drivers (connectors) for ODBC and JDBC—a critical part of our vertically integrated product portfolio—address the growing data governance challenges posed by existing applications and rapidly emerging AI agent variants. The @NotebookLM presentation below does an excellent job of explaining the problem (which continues to grow) and how our Request Broker and Database Agents have consistently addressed it. Enjoy!

Understanding our powerful Request Broker for orchestrating Database Agents (e.g., ODBC or JDBC Driver Bridges). It uses fine-grained, attribute-based access controls to govern connectivity to backend DBMS platforms.

Link to the OpenLink Request Broker Configurator used to manage various Database Agents, including ODBC- and JDBC-bridges i.e., ODBC to ODBC or ODBC to JDBC. github.com/OpenLinkSoftwa… You create and test Data Source Names (DSNs) using natural language via any SKILLS.md compliant AI Agent environment, that's it. High-performance and secure data connectivity based on existing open standards that will surpass any other the so-called native DBMS connectivity excuses racking up nothing but technical debt++

@Anthropic’s Claude Skills are a game changer in how software delivers units of economic value. They take us back to substantive code—Universal Filesystem Interface, Unix philosophy, Web principles, and more—without GUI detours or distractions. We (@OpenLink), we are successfully mapping key functionality from across our vertically integrated product portfolio into Skills that are usable wherever the SKILLS.md open standard is supported. For example, we now offer a Skill for configuring secure ODBC and JDBC connectivity—capabilities we’ve been refining since our founding in 1992, now ready for the agentic era. Our multi-tier edition drivers provide unique, fine-grained attribute-based access controls (ABAC) that protect data from the inevitable security pressures that will be introduced by AI agents. These controls govern ODBC and JDBC sessions using user, application, OS, network, and other contextual attributes—and even extend to third-party drivers (connectors).

Wondering why Virtuoso, our unique multi-model (tables and graphs) and multi-purpose (HTTP App Server & Filesystem), platform exists? Circa 2026, it still uniquely handles the real issues related to data access, integration, interoperability, and management.

@github repo link: github.com/OpenLinkSoftwa… #ODBC #DBMS #Database #Middleware

Here's our ODBC Driver Administrator Skill for creating and testing ODBC Data Source Names (DSNs). It's usable by any AI Agent and platform that supports the SKILLS.md open standard. Benefit? It makes ODBC easier to use, which means DBMS independence just got a lot simpler!

The power of Virtuoso is available via SKILLS.md compliant skills. Examples: [1] Virtuoso Support Agent -- collection of skills providing you with your own Virtuoso Expert Support Assistant [2] DataTwingler -- SQL, SPARQL, GraphQL query skills See: github.com/OpenLinkSoftwa…

Already there, thanks to what the Web offers gratis -- as the Internet's global filesystem layer. Usage examples, for collating AI Agent Skills from Github repos: [1] linkeddata.uriburner.com/describe/?url=… — Skills from our (@Openlink) Github Repo [2] linkeddata.uriburner.com/describe/?url=… — @garrytan's Skills collection [3] linkeddata.uriburner.com/describe/?url=… — @openclaw skills collection

@rom1504 @AnthropicAI @OpenLink @garrytan Screenshots, that I forgot to add.

The @OpenLink AI Layer (OPAL) provides a unique ability to create AI Agent Skills—using Agents.md and SKILLS.md—that are loosely coupled to Data Spaces (databases, knowledge bases, filesystems, and APIs). It achieves this by leveraging Virtuoso (#VirtuosoRDBMS) to bind to data spaces using existing open standards, making the entire process essentially a mapping exercise that outputs AI agent–readable markdown documents during the construction process. Most importantly, OPAL-generated skills are usable by any LLM or AI agent that understands SKILLS.md bundles. That’s it—the competitive moat is the unique ability to loosely couple data sources and services without compromising performance or governance-enforced security.

@OpenLink Here are documents generated by the Knowledge Graph generator AI Agent Skill mentioned in the slide deck. @GeminiApp produced the following, via @antigravity : #explore" target="_blank" rel="nofollow noopener">kingsley.idehen.net/DAV/home/kideh…

@OpenLink @GeminiApp @antigravity Variation of the same thing, but using @openAI's Codex UI and GPT-5.2. #skill-query-wikidata" target="_blank" rel="nofollow noopener">kingsley.idehen.net/DAV/home/kideh…

@OpenLink @GeminiApp @antigravity @OpenAI Ditto, but using @AnthropicAI's Claude Desktop and Sonnet 4.6. kingsley.idehen.net/DAV/home/kideh…

@aakashgupta Nice analysis. Here's it captured in an AI Agent friendly knowledge graph. See: linkeddata.uriburner.com/describe/?url=… Data Source (i.e., notes generated from your post). x.com/i/grok?convers… /cc @asemota

@Scobleizer @moltbook Here's your analysis in an AI Agent friendly knowledge graph. linkeddata.uriburner.com/describe/?url=… /cc @asemota

@asemota AI Agent friendly rendition of your insightful analysis via a publicly accessible Knowledge Graph. linkeddata.uriburner.com/describe/?url=… Raw notes: x.com/i/grok?convers…

@JasonSCui Here’s an AI Agent–friendly analysis of your insightful post, courtesy of @grok functioning as generic tool for crafting knowledge graphs. Naturally, this also provides rich context in machine-computable form for any AI Agent powered by an LLM. linkeddata.uriburner.com/describe/?url=…

"One information revolution having multiple successive applications." Absolutely Golden, from this latest @dwarkesh_sp podcast. These podcasts are so insightful!