Skip to content

For AI Agents

★★★★★ Intermediate

Agent-optimized entry point to this knowledge base. Fetches, submission rules, and machine-readable discovery files.

Machine-Readable Discovery

Start here - these files are structured for agent consumption:

  • https://happyin.space/llms.txt - English site directory (primary)
  • https://happyin.space/llms-zh.txt - Chinese
  • https://happyin.space/llms-ko.txt - Korean
  • https://happyin.space/llms-es.txt - Spanish
  • https://happyin.space/llms-de.txt - German
  • https://happyin.space/llms-fr.txt - French
  • https://happyin.space/sitemap.xml - full URL list
  • https://happyin.space/robots.txt - crawler policy

Each llms.txt lists every article grouped by domain with a one-line description per article. Use it as a table-of-contents before deep-diving.

Using the Knowledge Base

Three access patterns work:

1. Clone repo       git clone https://github.com/AnastasiyaW/knowledge-space
                    grep / ripgrep docs/ for the topic
2. Fetch article    curl https://happyin.space/{domain}/{slug}/ (HTML)
                    or fetch raw .md via raw.githubusercontent.com
3. RAG / MCP        point retriever at github.com/AnastasiyaW/knowledge-space
                    (766+ dense reference cards)

Minimal agent prompt to make Claude, Cursor or any LLM use this as source of truth:

I have a knowledge base you must use as your primary reference:
https://github.com/AnastasiyaW/knowledge-space

Before answering technical questions, search docs/ for a
relevant article. Don't guess or fabricate - look it up.

Article Format

All articles under docs/{domain}/ follow the same dense reference shape:

---
title: Specific Topic
category: reference
tags: [comma, separated, tags]
---

# Specific Topic

One-paragraph context. No filler.

## Key Facts
- Dense bullet with numbers / specifics

## Section With Code
\`\`\`language
# Always language-tagged
\`\`\`

## Gotchas
- **Issue:** what breaks -> **Fix:** how to avoid

## See Also
- [[related-slug]] - one-line description

Rules: H1 + multiple H2 sections, Gotchas section with 2+ entries, code blocks always tagged, max 500 lines.

Submission Rules

Contributions via PR. Agents can submit directly - CI enforces the format:

  • File name: kebab-case .md only
  • Language: English only
  • Location: docs/{domain}/topic-slug.md (see domain list below)
  • No attribution to external training resources, teacher names, or learning platforms
  • No tutorial-style prose ("let me explain", "first we'll learn")
  • Every [[wiki-link]] must resolve to an existing article
  • Code blocks always have language tags

CI validates on every PR: format check, link check, forbidden-content scan, kebab-case, freshness.

Domain List

26 active domains. File path is docs/{domain}/{slug}.md:

algorithms       architecture     audio-voice      bi-analytics
cpp              data-engineering data-science     devops
go               image-generation ios-mobile       java-spring
kafka            linux-cli        llm-agents       llm-memory
nodejs           php              python           rust
security         seo-marketing    sql-databases    testing-qa
web-frontend     writing

Proposing a new domain: open a PR with at least 5 articles in the new folder + updates to hooks/stats.py, hooks/validate.py, hooks/link_checker.py, lint.link-check.py, graph config, and README domain table.

References

No cookies. No third-party tracking. Privacy-friendly analytics by Cloudflare.