[PoC] aitools: Add SDK documentation query tool for MCP server #4261
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Add a new MCP tool `databricks_query_sdk_docs` that allows LLM agents to search Databricks SDK documentation for methods, types, and examples. This addresses the problem where LLMs struggle with the Databricks SDK because they lack indexed documentation. Instead of guessing API calls, agents can now query for proper method signatures, parameters, and usage. Features: - Fuzzy/keyword search across services, methods, types, and enums - Category and service filtering - Score-based result ranking - LLM-friendly markdown output with signatures and examples Implementation: - New sdkdocs provider with embedded JSON documentation index - Index generator tool that parses annotations_openapi.yml - Generated index includes 7 core services, 277 types, and 3 enums - Full unit test coverage for search and index loading Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
keugenek
requested review from
a team,
andrewnester,
anton-107,
denik,
lennartkats-db,
pietern and
shreyas-goenka
as code owners
keugenek
temporarily deployed
to
test-trigger-is
GitHub Actions
Inactive
Collaborator
|
Commit: 7a2f2ac
24 interesting tests: 22 KNOWN, 1 SKIP, 1 RECOVERED
Top 28 slowest tests (at least 2 minutes):
|
keugenek
temporarily deployed
to
test-trigger-is
GitHub Actions
Inactive
Adds a skill that helps Claude Code users discover and use the databricks_query_sdk_docs MCP tool effectively when asking about SDK methods, types, and parameters. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
keugenek
temporarily deployed
to
test-trigger-is
GitHub Actions
Inactive
keugenek
changed the title
Contributor
|
@keugenek does it have to be an MCP tools? I'd prefer this to be a command under |
keugenek
temporarily deployed
to
test-trigger-is
GitHub Actions
Inactive
This commit enhances the SDK documentation generator to parse the actual Go SDK source code instead of using hardcoded service definitions. Changes: - Rewrote tools/gen_sdk_docs_index.go to parse SDK using go/ast - Extracts service interfaces, method signatures, and descriptions - Parses struct types and enums automatically - Added tools/verify_sdk_docs_index.py for CI staleness check - Added Makefile targets: sdk-docs-index and verify-sdk-docs-index Results: - Previous: 7 services, 277 types, 3 enums (mostly hardcoded) - Now: 11 services, 1302 types, 263 enums (auto-generated) Usage: - make sdk-docs-index # Regenerate index - make verify-sdk-docs-index # Check if index is up to date Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
keugenek
temporarily deployed
to
test-trigger-is
GitHub Actions
Inactive
This commit adds GitHub Actions workflows to automatically manage the SDK documentation index: 1. check.yml: Added verification step to fail PRs with stale index - Runs `make verify-sdk-docs-index` on every PR 2. update-sdk-docs.yml: New workflow for automatic updates - Triggers on: manual dispatch, daily schedule, go.mod changes - Auto-commits to main when SDK version changes via push - Creates PR for scheduled/manual triggers if changes detected - Includes SDK version in commit messages This ensures the SDK docs index stays up to date when: - Dependabot bumps the SDK version - Manual SDK updates are made - Daily scheduled checks detect drift Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
keugenek
temporarily deployed
to
test-trigger-is
GitHub Actions
Inactive
keugenek
temporarily deployed
to
test-trigger-is
GitHub Actions
Inactive
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary
databricks_query_sdk_docsthat allows LLM agents to search Databricks SDK documentationrequests.get(...)instead of proper SDK methods)annotations_openapi.ymlwith 7 core services, 277 types, and 3 enumsFeatures
Example Usage
Test plan
go test ./experimental/aitools/lib/providers/sdkdocs/...make lintgo build -o cli .Related
Discussion in #db-agent-builder about APX MCP providing SDK documentation indexing for better LLM interactions.
🤖 Generated with Claude Code