Menu
Private knowledge base for indexing multimodal files or folders into a knowledge graph, supporting multi-hop graph retrieval
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
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)