SkillFoxx

RAGU plugin

The ragu-build plugin interviews you and assembles a graph RAG setup: a knowledge graph from text and search over it.

  • 145 ★ on the repository
  • Last commit
  • MIT
Plugin

Extended access

Why this level

  • The generated script runs extraction through a model and may send your documents to an external API.
All reasons and checks

raguteam/raguAuthor:RaguTeam

Install

In your terminal, with SkillFoxx CLI

npx skillfoxx add plugins/ragu

Detects the agents on your machine, shows the access level and pins the version.

Other ways to install

Access and checks

How we review

Extended accessThe tool runs code or makes network calls and only reads external data.

  • The generated script runs extraction through a model and may send your documents to an external API.

This is third-party code. Review the repository files before installing.

What it does

RAGU (Retrieval-Augmented Graph Utility) builds a knowledge graph from raw text and answers questions over it. The ragu-build plugin in this repository runs a short interview: the agent first gathers facts about the project and corpus (file extensions, file count, size), then selects a suitable RAGU pipeline. It outputs two files: `ragu_build.yaml` with the decisions and rationale, and a working `build_<name>.py` script. The script is checked with `scripts/validate_build.py`. The agent works from a module map and reads only the READMEs it needs rather than library source.

Who it is for. Developers and researchers building question-answering systems over their own documents who do not want to pick the chunker, extractor and storage by hand.

Good fit when

  • A folder of documents exists and you need to choose between local, global and mixed search.
  • A project starts with a knowledge graph over Russian-language texts where the NEREL ontology helps.
  • The agent should return a runnable build script and record why each component was chosen.
  • A team compares graph and vector storage options and records the decision in YAML.

Not a fit when

  • You need plain vector chunk search without a graph: simpler RAG libraries fit better.
  • The corpus is tiny and a knowledge graph will not repay entity extraction cost.
  • You have no model access for extraction and embeddings, locally or through an API.

Example request

Interview me about the document corpus in `docs/` and assemble a RAGU configuration, then generate `build_docs.py` and check it with the validator.

Limitations

The plugin assembles configuration and a script, while the `graph_ragu` library that actually builds the graph is installed separately and needs Python 3.10 or newer. Entity and relation extraction uses a language model through an API or local models (`pip install graph_ragu[local]`), which costs tokens and time. The NEREL ontology has 29 entity types and 49 relation types. The README has no install command for the plugin: connect its `.agents/skills/ragu-build` folder manually or through the repository marketplace file.

How to disable. Remove the `ragu-build` folder from the agent skills directory or disable the plugin with agent tooling. Uninstall the library with your package manager and keep or delete generated `ragu_build.yaml` and scripts by hand.

FAQ

What does the ragu-build plugin output?

Two files: `ragu_build.yaml` with decisions and rationale and a working `build_<name>.py`. The script is checked by a validator from the skill folder.

How do I install the library itself?

With `pip install graph_ragu`. For local transformers models there is `pip install graph_ragu[local]`.

In what language does the interview run?

In the language the user writes in. File names and code stay in English.

Which search types does RAGU have?

Local over graph neighborhoods, global over community summaries, naive over chunk vectors and a mixed engine.

Is there Russian documentation?

Yes, overviews of components, the NEREL ontology and the RAGU-lm extraction model exist in English and Russian under `docs/`.

Discussion

No questions or reviews yet.

Plugins

Semantic code search, cross-session memory and shell command checks for OpenCode. Runs locally on Ollama, with MCP and CLI modes

PluginExtended access64

A Claude Code and Codex plugin with a local RuvNet source base: the agent finds the right tool and shows the file with its code.

PluginExtended access83

The ak command installs, repairs and checks ruflo and agentic-qe, adds the offline RuvNet Brain base and links Claude Code, Codex

CLIExtended access61

A plugin with a local document and code index: hybrid semantic and keyword search, a dependency graph and three helper agents

PluginExtended access120

A plugin for spec-first development: six phases, subagent roles and a decision ledger, with a human approving each phase.

PluginExtended access87

A personal LLM wiki with a knowledge graph, cited RAG and memory, served to Claude Code and Cursor by an MCP server

MCP serverExtended access276

README in short

The library is built to move from a laptop to a server without rewrites: batched async work, rate-limited model requests and several graph and vector storages. The example shows graph construction with an OpenAI client, a chunker and a two-stage extractor.

README badge

Original SKILL.md

---

name: ragu-build
description: Interview the user about their RAGU use case, select an appropriate RAGU pipeline, and generate a validated ragu_build.yaml plus a runnable build_<name>.py script.

---

# Assembling a RAGU build

Goal: determine what the user actually needs through a short interview, then produce two artifacts:

* `ragu_build.yaml` — the recorded decisions and rationale;
* `build_<name>.py` — a working RAGU build script.

The RAGU library is already documented: every module ships a `README.md` describing its role in the pipeline and providing examples.

This skill is a **decision process and navigation map**, not a retelling of that documentation.

**Do not read module sources unless explicitly allowed below. Do not summarize whole READMEs.**
Read narrowly, guided by `references/module-map.md`.

Conduct the interview in whatever language the user writes in.

## Skill resources

All paths under:

* `references/`
* `assets/`
* `scripts/`

are relative to **this skill's own directory**, not to the user's project root.

Resolve the skill directory using whatever mechanism the current agent environment provides. Never assume that the skill lives inside the user's repository or that its resources are reachable through paths relative to the current working directory.

---

# Phase 0. Inventory

Before asking the user anything, infer everything that can reasonably be discovered from the project.

This removes unnecessary interview questions.

## 0.1 Read the module map

Report a problem