Agent Skills Hub

BattleScope

Streamline technical documentation for BattleScope features, maintaining consistency across API, frontend, and architecture layers.

cross-plugin-detective

Standardized detective skill integration for agent roles. Maps agents to code-analysis skills and enforces claudemem usage for memory-indexed code investigation.

rag-engineer

Architect and optimize production-grade RAG systems. Master embedding models, vector databases, chunking strategies, and retrieval pipelines for high-accuracy LLM applications.

scope-options

Plan mode on steroids. Push engineers to think with a product mindset before building with structured intake and concrete technical options.

nav-task

Manage Navigator task documentation: create implementation plans, archive completed features, and maintain the task index for Claude Code workflows.

project-overview

Complete project architecture and structure guide for LobeHub. Use for codebase exploration, project organization, file location, and architectural context.

seaborn

Statistical visualization library for Python. Create publication-quality graphics like box plots, heatmaps, and violin plots with pandas integration and automatic statistical estimation.

newsletter-events-write

Generate professional markdown newsletters from localized events stored in SQLite. Automates event aggregation, curation, and formatting for community or niche media newsletters.

swarm

Orchestrate parallel Claude Code worker swarms with protocol-based behavioral governance for complex features, multi-step refactors, and long-running autonomous coding sessions.

sql-toolkit

Command-line toolkit for SQL database management: schema design, query optimization, migrations, and performance debugging for SQLite, PostgreSQL, and MySQL.

agent-architecture

Implement production-grade AI agents with LangGraph, tool-calling guardrails, SSE streaming, and episodic memory. Includes anti-patterns, fix pairs, and stateful architecture patterns.

guard

Epistemic safety analysis for JSON data in prompts to prevent LLM hallucinations and reasoning errors when handling incomplete or large-scale datasets.