Standard.Agents.Data.Knowledge.MsSql 0.1.0

Standard.Agents.Data.Knowledge.MsSql

A SQL Server knowledge source for The Standard AI Agent Framework.

Part of The Standard for Agents. In Agent = Orchestration(Data, Decision, Direction), this package extends the Data nature — it moves knowledge from local files into a SQL Server table with full-text search. It is an optional adapter: it plugs into the core Standard.Agents package through the same IKnowledgeBroker seam, so nothing else about your agent changes.

The core Standard.Agents package ships file-based knowledge and is deliberately dependency-free. This package brings Microsoft.Data.SqlClient so you can ground an agent on a SQL Server table instead — with tokenized full-text search (FREETEXT) rather than the file default's substring match.

Install

dotnet add package Standard.Agents                       # the core framework
dotnet add package Standard.Agents.Data.Knowledge.MsSql  # this knowledge source

Use

using Standard.Agents;
using Standard.Agents.Data.Knowledge.MsSql;

var agent = new StandardAgent(apiUrl, apiKey, "LLooMA2.0")
    .UseKnowledgeMsSql("Server=localhost;Database=agent;Trusted_Connection=True;Encrypt=False;");

string answer = await agent.ProcessPromptAsync("What is our refund window?");

One line — everything else (skills, tools, guardians, memory, streaming) is the core, unchanged. Each turn the agent calls SelectKnowledgeAsync with the prompt; this broker runs a full-text query and returns the top matches, seeded into the agent's observations.

Options

.UseKnowledgeMsSql(
    connectionString,
    table: "KnowledgeDocuments",   // optionally schema-qualified, e.g. "kb.Documents"
    contentColumn: "Content",      // the text column searched and returned
    maxResults: 3)                 // documents injected per turn

Or construct the broker directly and pass it to the core's .UseKnowledge(...):

.UseKnowledge(new MsSqlKnowledgeBroker(connectionString, table: "kb.Documents"))

The table

This broker reads an existing table — populating it is your ETL, kept out of the agent. FREETEXT requires a full-text index on the content column:

CREATE TABLE KnowledgeDocuments (
    Id      bigint IDENTITY(1,1) PRIMARY KEY,
    Content nvarchar(max) NOT NULL
);

-- full-text search prerequisites (once per database / table)
CREATE FULLTEXT CATALOG AgentKnowledgeCatalog;
CREATE FULLTEXT INDEX ON KnowledgeDocuments (Content)
    KEY INDEX PK__KnowledgeDocuments   -- your PK's index name
    ON AgentKnowledgeCatalog;

The query the broker runs is, in effect:

SELECT TOP (@maxResults) Content FROM KnowledgeDocuments
WHERE FREETEXT(Content, @prompt);

FREETEXT matches on meaning — stemming and stop-words included — so "refund window" finds "windows for refunds." The prompt is always a bound parameter; the table and column names are your configuration, resolved once at construction. (Results are the full-text matches capped at maxResults; ranked ordering is a later enhancement.)

Where this fits

License

Licensed under the The Standard Software License (TSSL) v1.1 — the same license as Standard.Agents. Microsoft.Data.SqlClient is MIT-licensed and consumed as a NuGet dependency.