How do I save new knowledge to the Sparks registry?

Use the /learn command followed by a description of what you've discovered; the plugin will analyze the session context and propose a structured skill to save.

Does Sparks load knowledge automatically?

Yes, through SessionStart hooks, Sparks reads the project registry and injects relevant context instructions so Claude is aware of existing knowledge before you start working.

What categories of knowledge can I document?

Sparks supports specific categories including features, gotchas, patterns, decisions, procedures, performance, and testing strategy.

What is a 'Spark' in Claude Code?

A Spark is a project-scoped skill that represents captured knowledge, such as a debugging insight or an architectural decision, which Claude can automatically discover and load.

Sparks Knowledge Capture & Retrieval

Name: Sparks Knowledge Capture & Retrieval
Author: Codename-Inc

byCodename-Inc

学习与文档

Captures and retrieves durable project-specific knowledge to prevent context loss and build institutional memory within Claude Code.

关于

Sparks is a specialized knowledge management system for Claude Code that transforms ephemeral conversation insights into durable, project-scoped skills. By enabling the capture of hard-won debugging gotchas, architectural decisions, and feature dossiers, it ensures that critical context is preserved and automatically reloaded in future sessions. This plugin automates the documentation of 'why' and 'how' within your codebase, allowing both AI and human developers to leverage established patterns and avoid repeating past mistakes.

主要功能

Categorized knowledge registry covering architecture, gotchas, and patterns
SessionStart hooks for automatic injection of relevant registry data
0 GitHub stars
Context-aware searching and loading of project memory via /find
Automated knowledge extraction and skill generation via /learn command
Project-scoped storage that keeps documentation alongside the source code

使用场景

Creating searchable feature dossiers to accelerate onboarding for new team members or AI agents
Documenting architectural decision records (ADRs) and the rationale behind specific tech choices
Preserving 'gotchas' and debugging insights to prevent the same issue from being debugged twice

关于

主要功能

Categorized knowledge registry covering architecture, gotchas, and patterns
SessionStart hooks for automatic injection of relevant registry data
0 GitHub stars
Context-aware searching and loading of project memory via /find
Automated knowledge extraction and skill generation via /learn command
Project-scoped storage that keeps documentation alongside the source code

使用场景

Creating searchable feature dossiers to accelerate onboarding for new team members or AI agents
Documenting architectural decision records (ADRs) and the rationale behind specific tech choices
Preserving 'gotchas' and debugging insights to prevent the same issue from being debugged twice