ultra-rag
Routes UltraRAG pipeline orchestration, MCP server workflows, and the UI/backend stack.
- License 许可
- Apache-2.0
- 最近更新
- 2026-10-08
- Stars
- ★ 11
- Forks
- 3
01 前言 Front Matter
- name
- ultra-rag
- license
- Apache 2.0
- metadata
- { "disco-role": "operating" }
- description
- Routes UltraRAG pipeline orchestration, MCP server workflows, and the UI/backend stack.
- disable-model-invocation
- true
02 资源文件 4
- 参考文档references/overview.md4.2 KB
- 参考文档references/repo-provenance.md1.4 KB
- 参考文档references/repo-routing-metadata.json497 B
- 参考文档references/troubleshooting.md3.8 KB
03 SKILL.md 正文
UltraRAG
Use this repo skill when a task mentions UltraRAG pipelines, MCP servers, the web UI, the knowledge-base backend, or the case-study viewer.
UltraRAG is organized around three user-facing workflow families:
- Pipeline orchestration —
ultrarag build,ultrarag run,ultrarag show, YAML pipelines,ToolCall,PipelineCall, branching, looping, and the demo / experiment pipelines underexamples/. - MCP server workflows — the individual servers in
servers/, their tool and prompt signatures, parameter files, and backend-specific configuration. - UI and storage workflows — the Flask backend under
ui/backend, chat sessions, auth, knowledge-base management, memory sync, and the case-study viewer.
Start here
- Read
references/overview.mdfor the verified package shape, install hints, and repo map. - Read
references/repo-provenance.mdwhen you need to compare this skill to a checkout or decide whether to refresh it. - Read
references/troubleshooting.mdfor cross-cutting install, import, and runtime failures. - Use
scripts/check_install.pyfor a quick local import check.
Install and smoke check
Recommended install path:
uv sync
Narrower dependency sets are available when you only need part of the repo:
uv sync --extra retriever
uv sync --extra generation
uv sync --extra corpus
uv sync --extra evaluation
uv sync --all-extras
Editable installs also work:
uv pip install -e .
uv pip install -e '.[all]'
Verified package facts:
- Distribution:
ultrarag - Version:
0.3.0.2 - Python:
>=3.11, <3.13 - Console entry point:
ultrarag = ultrarag.client:main
Minimal check:
python -c "from importlib.metadata import version; print(version('ultrarag'))"
ultrarag --help
Route map
| If the task is about... | Read |
|---|---|
ultrarag build, ultrarag run, ultrarag show, ToolCall, PipelineCall, YAML step flow, branch/loop logic, example pipelines |
sub-skills/pipelines/SKILL.md |
servers/retriever, servers/generation, servers/prompt, servers/corpus, servers/evaluation, servers/custom, servers/memory, backend options, tool/prompt signatures, parameter files |
sub-skills/servers/SKILL.md |
ultrarag show ui, ultrarag show case, the Flask backend, chat sessions, KB files, auth, memory sync, storage paths, frontend build/layout |
sub-skills/ui-and-storage/SKILL.md |
Shared runtime helpers
scripts/check_install.py— quick import/version check for the installed package.
Notes for future agents
- Keep workflow instructions inside the generated skill tree; do not depend on the original checkout being open.
- Use the sub-skill that matches the user intent first, then open the nearest reference file for command details, backend notes, and troubleshooting.
- If a task spans multiple families, read the overlapping sub-skills instead of guessing from source layout.
04 同仓库更多 Skills
obsidian-clipper-template-creator
Guide for creating templates for the Obsidian Web Clipper. Use when you want to create a new clipping template, understand available variables, or format clipped content.
obsidian-clipper-template-creator
Guide for creating templates for the Obsidian Web Clipper. Use when you want to create a new clipping template, understand available variables, or format clipped content.
python-api
Programmatic Object-Detection-Metrics API guidance for BoundingBox, BoundingBoxes, Evaluator, enums, coordinate conversion, AP/mAP semantics, and source-style import diagnostics.
file-evaluation
Folder-based PASCAL VOC AP/mAP evaluation from ground-truth and detection text files.
object-detection-metrics
Guides legacy Object-Detection-Metrics PASCAL VOC AP/mAP evaluation from detection text folders and direct Python API classes.
nx-workspace-patterns
Configure and optimize Nx monorepo workspaces. Use when setting up Nx, configuring project boundaries, optimizing build caching, or implementing affected commands.