GitHub·· 9 天前AI 评分20
joanmarcriera/Home-office-automations：面向隐私优先家庭办公自动化的智能体维护知识库
joanmarcriera/Home-office-automations
AI 导读
GitHub 项目 joanmarcriera/Home-office-automations 是一个由 LLM 智能体维护的结构化知识库，用于搭建和运营隐私优先的家庭办公自动化栈，配套文档站点 ai.riera.co.uk。
正文
A structured, agent-maintained knowledge repository for building and operating a privacy-first home-office automation stack.
🌐 Live Documentation: ai.riera.co.uk
🔍 What You Will Find Here
- Canonical docs for AI tools, self-hosted services, and orchestration patterns.
- Practical playbooks for recurring workflows.
- Architecture maps and data-flow documentation.
- Automated intake and maintenance workflows that continuously expand the knowledge base.
🗺️ Repository Map
| Section | Path | What it contains | Start here if you need to... |
|---|---|---|---|
| Documentation home | docs/index.md |
The MkDocs landing page and high-level navigation. | Understand the full docs structure quickly. |
| Tool catalogue | docs/tools/ |
Canonical pages for AI tools, frameworks, providers, agents, infra, and benchmarking. | Compare tooling options or add a new tool page. |
| Self-hosted services | docs/services/ |
Service-level docs for the homelab stack (storage, automation, media, networking). | Deploy or evaluate a self-hosted component. |
| Knowledge base | docs/knowledge_base/ |
Cross-cutting concepts: MCP/ACP, model classes, RAG, security, and landscape overviews. | Learn the underlying concepts before implementation. |
| Playbooks | docs/playbooks/ |
Step-by-step operational runbooks. | Execute a concrete workflow end-to-end. |
| Reference implementations | docs/reference-implementations/ |
Reusable prompts, mapping rules, and workflow exports. | Reuse proven templates/config patterns. |
| Architecture | docs/architecture/ |
Component map, flow diagrams, governance, and automation model. | Understand how systems connect and why. |
| Intake queue | docs/new-sources.md |
Source ingestion process plus dated logs under docs/new-sources/. |
Track what is new, pending, and integrated. |
| Standards | docs/standards.md |
Taxonomy, canonical-page rules, and documentation contract. | Validate structure before opening a PR. |
| Contribution workflow | docs/CONTRIBUTING.md |
Human + agent contribution requirements and quality gates. | Contribute safely without breaking repo contracts. |
| Agent operating guide | AGENTS.md |
Durable rules/checklists for autonomous LLM agents working in this repo. | Run maintenance work with consistent outputs. |
| Agent skills guide | skills.md |
Reusable task patterns for intake, docs updates, and quality checks. | Execute recurring tasks with low variance. |
🚦 Recommended Entry Points
- New maintainer:
docs/index.md - New contributor:
docs/CONTRIBUTING.md - LLM/automation agent:
AGENTS.mdthenskills.md - Architecture review:
docs/architecture/README.md - Tool selection:
docs/knowledge_base/ai_tooling_landscape.md
🛠️ Utility Commands
# List all known tools/services from the catalog python3 scripts/compare_agents.py list # Open the canonical documentation path for a tool id python3 scripts/compare_agents.py doc paperless-ngx # Validate catalog/nav consistency before a docs PR python3 scripts/check_catalog_consistency.py
🎯 Project Scope
This repository is optimized for long-term maintainability:
- One canonical page per tool/framework/provider.
- Cross-linked documentation with stable taxonomy.
- CI quality gates for catalog consistency and docs contract compliance.
- Scheduled GitHub workflows for source intake and backlog processing.\n
Tracking
Vikunja project 72 · Home-office-automations — https://familia.riera.co.uk/projects/72
来源：GitHub · github.com