Knowledge Discovery FAQs

Question 1

When should I use this skill?

Accepted Answer

You should invoke this skill at the start of any non-trivial task, such as implementing new features, debugging complex errors, or exploring a new repository. It is especially useful when you need to know which specific skills (like EFCore or ABP patterns) should be loaded for a task.

Question 2

How does the tiered discovery model improve efficiency?

Accepted Answer

The model optimizes performance by matching discovery depth to task complexity. It ranges from 'SKIP' for simple typos (0s cost) to 'DEEP' for full feature implementations (15-20s cost), ensuring Claude doesn't waste time reading files for obvious tasks while remaining thorough for complex ones.

Question 3

How does Knowledge Discovery handle skill dependencies?

Accepted Answer

The skill uses a tiered 'Layer' system and a context graph to ensure skills are loaded in the correct order (e.g., loading Foundations before Frameworks). This prevents logic gaps and ensures that Claude applies architectural patterns consistently across the project.

Question 4

Can this skill help with framework-specific patterns?

Accepted Answer

Yes. It includes built-in quick-reference lookups for popular frameworks and libraries including ABP Framework, EFCore, OpenIddict, and FluentValidation. It can instantly map common error messages to the specific technical skills required to fix them.

Question 5

What does the Knowledge Discovery skill do?

Accepted Answer

Knowledge Discovery acts as an intelligent navigator for complex codebases. It uses a tiered efficiency model to identify relevant technical patterns, dependencies, and workflows, ensuring Claude has the exact context needed for any development task without unnecessary overhead.

Knowledge Discovery

Knowledge Discovery

概要

主な機能

ユースケース

概要

主な機能

ユースケース