How do I add new knowledge to the Lore database?

You can use the 'lore add' command followed by a title, type (like convention or gotcha), and the markdown body to preserve information for yourself and the AI agent.

What happens if I try to query a project that doesn't have Lore?

The skill includes a 'lore doctor' check. If the .lore directory is missing, Claude will inform you that the project isn't set up yet and suggest running 'lore init' to begin.

How does Lore differ from a standard documentation file?

Unlike standard markdown files that provide all information at once, Lore provides scoped, queryable entries that return only what is relevant to your specific task or file path, preventing the AI from reaching its context limit.

Can I use Lore to document architectural decisions?

Yes, Lore includes specific entry types for 'architecture' and 'decision', making it an ideal tool for maintaining lightweight Architecture Decision Records (ADRs) that the AI can reference.

Lore Project Knowledge

Name: Lore Project Knowledge
Author: namishsaxena

bynamishsaxena

•

Productivity & Workflow

Manages and retrieves project-specific conventions, architecture decisions, and code patterns through scoped AI memory.

Lore Knowledge acts as a specialized persistent memory for your repository, allowing Claude to access only the most relevant architectural decisions, code patterns, and project 'gotchas' without context window bloat. By replacing large, monolithic documentation files with a queryable database, this skill ensures that your AI agent understands your project's unique domain context and coding standards while keeping conversations focused and efficient. It enables Claude to proactively learn from the codebase and preserve tribal knowledge for future tasks.

Key Features

01Categorized entry types including gotchas, conventions, and ADR-style decisions

02Knowledge capture system to preserve newly discovered patterns and architectural decisions

03Task-based knowledge retrieval to find relevant patterns for specific implementation goals

042 GitHub stars

05File-scoped context discovery to understand conventions in unfamiliar directories

06Health monitoring via the doctor command to ensure local knowledge availability

Use Cases

01Reducing token usage and costs by providing Claude only with specific project lore relevant to the current task

02Onboarding to an unfamiliar codebase by querying existing patterns and common pitfalls

03Maintaining architectural consistency by retrieving project-specific conventions before a refactor

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add namishsaxena/lore lore-knowledge

For use in Claude.ai and ChatGPT

Download Skill