Sign InOpen Brain
AI EngineerVideoSource Linked

Rebuilding the web for agents — Liad Yosef, MCP Apps

Publishing agent instructions is not enough: tested agents mostly ignored llms.txt and navigated documentation directly, so agent readiness needs observed journey testing and discoverable interfaces.

AI Engineer · Sep 16, 2026
Open Source Open MarkdownOpen JSON
Source Summary

Aura found that **almost 50% of tested sites** published llms.txt, yet almost none of the tested agents used it without another page pointing them there. Agents generally went first to documentation and then the homepage.

Practical Implication

Test real tasks across several agent harnesses and inspect their paths instead of treating one metadata file as compliance. Expose callable resources through standards such as **MCP Apps** and **AI catalog.json**, while retaining UI for choices that still require human judgment.

Agent-Ready Context
Aura found that **almost 50% of tested sites** published llms.txt, yet almost none of the tested agents used it without another page pointing them there. Agents generally went first to documentation and then the homepage.

Test real tasks across several agent harnesses and inspect their paths instead of treating one metadata file as compliance. Expose callable resources through standards such as **MCP Apps** and **AI catalog.json**, while retaining UI for choices that still require human judgment.

The findings come from Aura’s own scans and journey tooling; the talk does not disclose the sampled sites, agents, or scoring method in enough detail for independent comparison. The broader prediction of a nearly headless web is directional, not established.
Connected Context · Feed7 Judgment

This narrows agent-ready web design: publishing agent metadata is ineffective if harnesses do not discover or consult it. The practical requirement becomes journey testing across agents plus callable, standardized resources, with UI preserved where people must judge choices. It shifts emphasis from nominal compliance to observed access paths, while leaving the headless-web prediction unproven.

MCP Apps: Extending the Frontier — Ido Salomon & Liad YosefMCP Apps supplies the concrete host-governed UI mechanism for the Signal’s recommendation to expose callable resources without removing human-facing controls.ChromeDevTools/chrome-devtools-mcpChrome DevTools MCP provides the runtime evidence needed to inspect real agent journeys—navigation, network activity, console output, and screenshots—instead of assuming an agent-facing file is discovered.The Dark Arts of Web Automation: Teaching Agents to Use Websites Like Humans — Corey Gallon, RexmoreBrowser agents can still traverse conventional sites through verified sense-act loops; this Signal argues that standardized callable resources can reduce dependence on such brittle navigation while retaining UI for judgment-heavy steps.Latest MCP spec now supported in mcp-handlerAdopting MCP as an agent-facing access layer creates an implementation consequence: servers must account for current transport, runtime, and SDK compatibility rather than treating protocol publication alone as sufficient.
Context Map
contextcoding#mcp#tool-use#computer-use
Uncertainty
The findings come from Aura’s own scans and journey tooling; the talk does not disclose the sampled sites, agents, or scoring method in enough detail for independent comparison. The broader prediction of a nearly headless web is directional, not established.