feat: wire claude-code agent runner backend

This commit is contained in:
Andrey Avtomonov 2026-05-15 13:04:16 +02:00
parent eb90d2f32c
commit 3de32c43a1
16 changed files with 229 additions and 21 deletions

View file

@ -250,6 +250,17 @@ canonical revenue reporting.
Together, these four pillars give agents enough context to produce analytics artifacts that match what your team would produce - not just syntactically valid SQL, but the right query for the question.
## Agent runner backends
KTX separates the global LLM provider from the agent runner. The global
provider powers non-agent calls such as scan enrichment and relationship
proposals. The agent runner powers curated tool loops used by ingest and memory
capture.
Use `llm.agentRunner.backend: claude-code` when you want those curated loops to
run through the Claude Agent SDK. KTX registers only its stage-specific MCP
tools for the session and disables Claude Code built-in tools for that backend.
## How KTX compares
KTX is a context layer with an agent-native semantic layer at its core. MetricFlow, Cube, and Malloy model metrics, dimensions, joins, and generated SQL. KTX covers that semantic-layer work, then adds the context agents need to use and maintain it: wiki pages, schema scans, provenance, ingestion, validation, and agent-facing CLI commands.