knowledge-reader

name	knowledge-reader
description	Navigate the pre-built Knowledge Base graph to discover repo conventions, architecture, testing patterns, and verification strategies before coding.
topics	knowledge base, KB, conventions, testing, architecture, patterns, verification, dependencies, code style, error handling

Knowledge Base Reader

You have access to a pre-built Knowledge Base (KB) for the target repository. The KB index path and KB directory are listed in the Pre-flight section of your prompt — read the KB index first to see what categories are available.

The KB is a structured graph of markdown documents organized into categories. Each category has an index file and detailed leaf files. Start by reading the KB index, then use the KB directory path to read specific leaf files relevant to your task.

When to Read KB Leaf Files

Consult the KB before you:

Situation	Read from these KB categories
Write or modify code	conventions/ — read the index, then leaf files for error handling, naming, code style
Write or modify tests	verification/ + conventions/ — test patterns, test infrastructure, testing conventions
Make architectural decisions	architecture/ — components, data flow, key abstractions
Use or extend APIs	api-surface/ — services, client libraries, config schemas
Add dependencies	dependencies/ — approved libraries, version constraints
Set up verification steps	verification/ — test suites, CI pipeline, verification order

How to Navigate

1. Read the KB index from the path listed in Pre-flight — it shows the category map, build commands, key architecture decisions, and important file paths
2. Identify which categories are relevant to your current task using the table above
3. Read the category's index.md first (e.g., <KB Directory>/conventions/index.md) to discover the actual leaf filenames — these vary per repo
4. Read the relevant leaf files based on what the category index lists
5. Only read what you need — typically 2-4 leaf files per task

Category Overview

Category	What it covers
architecture/	Component maps, data flow patterns, dependency injection modules, key abstractions, concurrency and persistence models
conventions/	Error handling patterns (including error grading), naming conventions, testing patterns, code style, logging and context propagation
api-surface/	gRPC services, CLI tools, client library API, configuration schemas, environment variables, Kafka event contracts
dependencies/	Go library inventory, build tools (Task, Bazel, etc.), runtime services (Cassandra, CRDB, Kafka), version constraints
verification/	Unit test patterns, integration test infrastructure, CI pipeline, linting rules, recommended verification order

Important Notes

• The KB may be slightly behind the current codebase. If KB content conflicts with what you see in the actual source code, trust the source code.