원클릭으로
multimodal-rag
Private knowledge base for indexing multimodal files or folders into a knowledge graph, supporting multi-hop graph retrieval
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
메뉴
Private knowledge base for indexing multimodal files or folders into a knowledge graph, supporting multi-hop graph retrieval
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Schedule reminders and recurring tasks. Actions: add, list, remove or set_context.
Periodic wake-up service that checks HEARTBEAT.md for pending tasks and executes them automatically. Use this skill to write or update HEARTBEAT.md.
Generate wiki docs + Mermaid diagrams for any codebase. Use when the user asks to document a codebase, generate architecture diagrams, or create a structured wiki for a repo.
Parse an image from a file path to obtain a description, enabling non-multimodal LLMs to have vision capabilities.
When the user needs to transcribe video (such as .mp4, .mkv, .avi) into text, use the python_repl tool to generate text.
Karpathy's LLM Wiki: build/query interlinked markdown KB.
| name | multimodal_rag |
| description | Private knowledge base for indexing multimodal files or folders into a knowledge graph, supporting multi-hop graph retrieval |
Index all files under a directory into the rag-anything graph:
import asyncio
from skills.builtin.core.multimodal_rag.scripts import folder_index, file_index
if __name__ == "__main__":
_classify_folder: str = "{placeholder}" # <- replace with the knowledge graph category
# Use this when the input is an entire folder
_input_folder_path: str = "{placeholder}" # <- replace with the absolute path of the input folder
coro = folder_index(_input_folder_path, _classify_folder)
# Use this when the input is a single file
# _input_file_path: str = "{placeholder}" # <- replace with the absolute path of the input file
# coro = file_index(_input_file_path, _classify_folder)
# Run
res = asyncio.run(coro)
print(res)
Query the rag-anything knowledge graph:
import asyncio
from skills.builtin.core.multimodal_rag.scripts import query
if __name__ == "__main__":
_query: str = "{placeholder}" # <- replace with the question
res = asyncio.run(query(_query))
print(res)