life.ayushrai.site product MCP server

Personal Knowledge Base (PKB)

This repository is a long-term personal knowledge system. The website is one optional consumer; not the source of truth.

Start here

1. docs/KNOWLEDGE_SYSTEM.md
2. docs/FILE_STRUCTURE.md
3. docs/ENTITY_MODEL.md
4. docs/RELATIONSHIP_MODEL.md
5. docs/CONTENT_WORKFLOW.md

Layout

Principles

• One fact, one place
• Reference by entity id; never duplicate
• Evidence over marketing language
• Uncertain identity data → /knowledge/review, never silent delete

Commands

npm run validate       # knowledge validation (Node)
npm run build:site     # generate resumes, portfolios, domains from data/site-catalog.json
npm run qa             # static QA (em dashes, viewport, catalog coverage); exits 1 on failure
npm run check:agent    # agent surface checks (negotiation, discovery, jobs, trust pages)
npm run local            # full local site on http://localhost:3000 (vercel dev: middleware, rewrites, api/)
npm start              # static-only preview (Python http.server; no rewrites or api/)
node scripts/build-home.js   # regenerate homepage from knowledge/_system/homepage.yaml
npm run build:collections    # work index, books, influences from knowledge/{work,books,influences}/*.json

Adding work, books, people, essays, and policy case studies: docs/workflows/ADDINGLIFECONTENT.md.

Browse the homepage at / under npm run local. Public site: https://life.ayushrai.site/. Developer portal: https://life.ayushrai.site/developers/.

Agent skills (skills.sh)

npx skills add ayushrai-hub/LifeOS

See skills/README.md.

SDK / CLI packages

• JavaScript: packages/life-ayushrai (life-ayushrai CLI)
• Python: packages/life_ayushrai

Publish with owner registry auth (npm publish / python -m build && twine upload).

Website consumer

The site is a knowledge graph in HTML: no mega-menu, inline blue links, multiple resumes and portfolios, domain pages, certifications, and media entities. Edit data/site-catalog.json, then run npm run build:site.

Deploy (Vercel)

• middleware.js rewrites / → sites/www/ and 308s /sites/www → /
• vercel.json maps root agent URLs to files in discovery/
• Content chronology lives in knowledge entities + knowledge/_system/site-status.yaml

npx vercel

Identity spreadsheet

Drop the master online-presence file into data/imports/ and run the import script.

Filling personal data

Provide biography facts, work dates, certifications, and resume details. Until then, entity sections stay empty on purpose; nothing is invented.