Agent Skills by msg
Showing 39 vetted skills indexed across 4 GitHub repositories.
daily-webnovel-writing-knowledge
Construct a knowledge graph for Korean webnovel writing with ontology and node/edge definitions.
spring-modern-api
Provides Spring 6 and Spring Boot 3 API development and implementation guidance.
codebase-understanding
Build a queryable codebase knowledge graph to analyze architecture and dependencies.
teddynote-langchain-rag
Troubleshoot RAG and LangChain pipelines covering loading, embedding, retrieval, and evaluation.
modern-java-in-action
Guide modern Java idioms for backend code with lambdas and streams.
text-knowledge-skill-builder
Converts text documents into structured knowledge graphs and reusable skills.
moonshotnote-ocr
Extract and verify Korean and English text from images using OCR engines.
fastapi-clean-architecture
Explore FastAPI and clean architecture concepts through a knowledge graph.
backend-architecture
Provides knowledge graphs for backend architecture design focusing on layering and dependency management.
moonshot-research
Execute public-web research queries across search engines, social media, and developer forums.
tidy-first
Query a knowledge graph for code tidying best practices and trade-offs.
moonshot-decide-sequence
Determine Moonshot workflow phases and execution chains from analysis context.
project-md-refresh
Generate and update PROJECT.md from repository structure and config files.
commit-moonshot
Analyze repository changes, update Memory MCP entities, and commit updates.
failure-analyzer
Identifies recurring failure patterns in agent runs and outputs improvement patches for PROJECT.md, CLAUDE.md rules, and SKILL.md files.
web-design-guidelines
Assesses UI code conformity with Web Interface Guidelines, reporting accessibility and design violations by file:line.
moonshot-evaluate-complexity
Evaluate task complexity from file, line, and time estimates.
moonshot-phase-runner
Automate phase-by-phase Moonshot execution from a master plan.
completion-verifier
Run acceptance tests and trigger retries on implementation failures.
moonshot-classify-task
Classify user requests into feature, modification, bugfix, or refactor task types.
moonshot-orchestrator
Analyze user requests and construct optimal agent chains for PM workflows.
doc-auto-sync
Detect code changes and update affected project documentation files.
build-error-resolver
Diagnose and fix TypeScript, lint, and module resolution build errors.
workflow-self-improver
Apply failure-analyzer improvements to CLAUDE prompts, rules, and settings.