mirror of
https://github.com/different-ai/openwork
synced 2026-04-25 17:15:34 +02:00
bd546147e5
* feat(landing): make site agents-ready Adds robots.txt with AI crawler rules and Content-Signal (ai-train=no, search=yes, ai-input=yes), an RFC-compliant sitemap.xml, Link headers pointing agents to /docs and the agent-skills index, and exposes the workspace-guide skill at /.well-known/agent-skills/ with a sha256-verified index.json. * feat(landing): permit AI training in Content-Signal * feat(landing): simplify robots.txt and add llms.txt navigation guide Collapse robots.txt to a single User-agent: * Allow: / block (matching the Huggingface pattern) — all per-bot entries had identical rules, so the wildcard already covers them and future bots without maintenance. Content-Signal remains permissive (ai-train=yes, search=yes, ai-input=yes). Add llms.txt per the llmstxt.org spec — the main gap Cloudflare's agent-readiness checklist identifies for this site. It tells agents how to route users based on intent (download, pricing, enterprise, den), what agents can do once OpenWork is installed, and how the cloud plan differs from the desktop app. * docs(landing): neutral llms.txt tone, drop feedback routing Drop the routing entry that sends agents to /feedback — feedback should come from humans, not agent-mediated submissions. Rephrase the agents note from 'we want usage' to a neutral statement of what agents may do.