Skill Claude CodeCodex
Use katok to dry-run, stage, or send one explicitly confirmed KakaoTalk message on macOS through the local Accessibility-driven KakaoTalk UI. Never use for bulk, repeated, unsolicited, deceptive, harassing, or unattended delivery.
Skill Claude CodeCodex
Search local KakaoTalk keyword, BM25, and EmbeddingGemma vector indexes through the katok CLI, list rooms, export chunks or transcripts, extract attachments, and prepare guarded message sends only after explicit user confirmation.
Skill Claude CodeCodex
Query AutoRAG-Research pipeline results using natural language. Converts questions to SQL, executes safely (SELECT-only), returns formatted results. Auto-detects DB connection from configs/db.yaml or env vars. Use for pipeline comparison, metrics analysis, token usage.
Skill Claude CodeCodex
Guide developers through creating a custom generation pipeline plugin for AutoRAG-Research. Walks through scaffolding, implementing BaseGenerationPipeline methods, composing with retrieval pipelines, writing YAML configs, testing, and installing. Use when building a new RAG generation strategy (e.g., chain-of-thought…
Skill Claude CodeCodex
Guide developers through creating a custom data ingestor plugin for AutoRAG-Research. Ingestors load external datasets (HuggingFace, local files, APIs) into the database. Uses @registeringestor decorator for automatic CLI parameter extraction. Use when ingesting a new dataset format into AutoRAG-Research.
Skill Claude CodeCodex
Guide developers through creating a custom evaluation metric plugin for AutoRAG-Research. Covers both retrieval metrics (recall, precision, etc.) and generation metrics (BLEU, ROUGE, etc.). Walks through scaffolding, implementing metric functions with @metric decorators, writing configs, testing, and installing. Use…
Skill Claude CodeCodex
Guide developers through creating a custom retrieval pipeline plugin for AutoRAG-Research. Walks through scaffolding, implementing BaseRetrievalPipeline methods, writing YAML configs, testing, and installing. Use when building a new search/retrieval strategy (e.g., Elasticsearch, ColBERT, custom vector search).
Agent Claude Code
Use this agent when you need to analyze and understand the structure of an external dataset (primarily from HuggingFace) before writing ingestion code. This agent inspects raw data structures, column names, data types, and nesting patterns without writing implementation code.\n\nExamples:\n\n \nContext: User wants to…
Agent Claude Code
Use this agent when you need to implement a new data ingestor class based on an architectural blueprint (MappingStrategy.md). This agent translates design specifications into production-ready Python code following the AutoRAG-Research service layer patterns. Examples of when to use this agent:\n\n \nContext: User has…
Agent Claude Code
Use this agent when you have a PipelineAnalysis.json and need to map the algorithm to existing patterns (base classes, services, UoW) in the AutoRAG-Research codebase. Context: Paper analysis is complete, now need to design the architecture. user: "Design the architecture for the HyDE pipeline" assistant: "I'll use…
Agent Claude Code
Use this agent when you have PipelineDesign.md and tests written, and need to implement the actual pipeline code to pass the tests. Context: Tests are written and failing, now need implementation. user: "Implement the HyDE pipeline to pass the tests" assistant: "I'll use the pipeline-implementer agent to write the…
Agent Claude Code
Use this agent when you need to analyze a research paper (PDF, arxiv link, or URL) to extract algorithm details for implementing a new RAG pipeline. Context: User wants to implement a new retrieval pipeline from a research paper. user: "Implement the HyDE retrieval pipeline from this paper…
Agent Claude Code
Use this agent when you have a PipelineDesign.md and need to write tests BEFORE implementation (TDD approach) for a new RAG pipeline. Context: Architecture design is approved, now need to write tests first. user: "Write tests for the HyDE pipeline" assistant: "I'll use the pipeline-test-writer agent to create…
Agent Claude Code
Use this agent when pipeline implementation is complete and you need to run validation (tests, type checks, linting) before committing. Context: Implementation is complete, need final validation. user: "Validate the HyDE pipeline implementation" assistant: "I'll use the pipeline-validator agent to run all checks…
Agent Claude Code
Use this agent when you need to design an ingestion strategy for mapping external data sources to the AutoRAG-Research PostgreSQL schema. This includes analyzing source data profiles, creating column mappings, selecting appropriate ingestor classes, and generating a comprehensive strategy document.\n\nExamples:\n\n…
Agent Claude Code
Use this agent when you need to generate test code for new or existing functions, classes, or modules. This includes when: (1) A new function or class has been implemented and needs corresponding tests, (2) Existing code lacks test coverage, (3) You want to verify behavior of recently written code, (4) Refactoring…
Command Claude Code
Scan recently modified Python files for utility function duplication.
Command Claude Code
Stage changes, commit with descriptive message, and push to current branch.
Command Claude Code
Design ingestion strategy with mandatory human review. Use after dataset inspection phase.
Command Claude Code
Orchestrate full pipeline implementation workflow from paper to validated code. Use when implementing a new retrieval or generation pipeline from a research paper.
Command Claude Code
Create a pull request to main branch.
Command Claude Code
3-agent PR code review debate -- spawns Devil's Advocate, Neutral Judge, and Approval Advocate to review current PR.
Command Claude Code
Process [APPROVE] and [IGNORE] replies on /refactor review threads -- applies approved fixes and resolves all responded threads.
Command Claude Code
PR summary comment -- what problem this PR solves, how it solves it, and key human-in-the-loop decisions.