Chinng Lab AI-Agent Portal
AI-agent-first content hub serving news, CVE advisories, and Japanese/US government documents metadata-first over MCP and REST, plus an Agent-Readiness directory scoring how well other sites expose agent surfaces.
AI Documentation Files
Access the llms.txt files for this website
An AI-agent-first content hub. Crawlers ingest articles three times a day and an asynchronous worker enriches them with LLM summaries, machine translation to English, and knowledge-graph entity lookup. Everything is served metadata-first — title, summary, entities, related links — so agents can filter before fetching bodies. Article bodies are gated per each item's redistribution license: first-party content is served in full, third-party sources are link-only.
Key Focus Areas
- News articles (
news/) — crawled and enriched withsummary_enas the English pivot - CVE and security advisories (
security/) — vulnerability feed with its own sitemap - Government documents (
government/) — Japanese agencies and the U.S. Federal Register, with provenance fields - Agent-Readiness directory — scores other domains 0–100 across Discoverability, Content, Bot Access Control, API/Auth/MCP/Skill, and Commerce, mapped to Level 0–3
- Citation Packs — per-article attribution, license, content hash, and reuse policy at
/citations/<id>.json
Agent Surfaces
- MCP server:
https://portal.chinng-lab-srv.dev/mcp(streamable-http, eight tools) - MCP Server Card:
/.well-known/mcp/server-card.json - AI catalog (Agentic Resource Discovery):
/.well-known/ai-catalog.json - API catalog (RFC 9727 linkset):
/.well-known/api-catalog - Agent Skills index:
/.well-known/agent-skills/index.json - REST, no auth:
/api/search,/api/government,/api/popular,/api/changes,/api/surfaces - Markdown views: every article is
text/markdownat its canonical URL; category feeds at/feeds/<category>.md
About llms.txt Implementation
The portal's llms.txt documents the full read surface — MCP tools, REST endpoints, the license model that governs body exposure, and the incremental-monitoring pattern via content_hash change detection — so an agent can go from discovery to a correctly attributed citation without scraping HTML. llms-full.txt expands this with the complete endpoint and schema detail. Sitemaps are split by content family (sitemap_news.xml, sitemap_cve.xml, sitemap_government.xml, sitemap_sites.xml) so agents never mix categories.
Developer Tools
Explore tools created to help you work with llms.txt
- LLMs.txt Checker Chrome ExtensionCheck if websites implement llms.txt and llms-full.txt files
- LLMS.txt VSCode ExtensionSearch and explore llms.txt files directly in VS Code
- MCP LLMS.txt ExplorerExplore and analyze llms.txt files using MCP
- LLMs Txt Raycast ExtensionSearch and explore llms.txt files directly in Raycast
- llmstxt CLIInstall llms.txt documentation directly into your AI coding agents
Browse more
Related Projects
Discover similar websites implementing llms.txt
A machine-readable directory where AI products register themselves so AI agents can discover them. Autonomous registration, free listings, JSON, schema.org.
LiveFaceSwap AI provides real-time browser face swap, try-on, and restyle previews, plus Windows 11 virtual-camera output through its desktop app.
One-click AI pipelines for video, image, audio, and text — runnable from the web, a CLI, SDKs, or an MCP server.