News
In short
15 dated changes to crawlers, agents and the vendors at the door, from 2026-08-19 to 2026-11-02, newest first. Each item names the rows in the Crawler Index it affects and links its operator or vendor source, read 2026-11-04. The latest, from 2026-11-02, is Signed-Agent-Metadata header proposed for Web Bot Auth.
- 2026-11-02Signed-Agent-Metadata header proposed for Web Bot Auth
draft-vilgalys-webbotauth-signed-agent-metadata-00 proposes an HTTP header that carries signed metadata about the agent alongside a Web Bot Auth signature. - 2026-10-29Vercel splits AI crawlers from AI assistants
Vercel's AI Bots managed ruleset can now log or block AI crawlers (indexing and training) and AI assistants (user-triggered fetches) independently, on all plans. - 2026-10-29Web Bot Auth working-group draft reaches -01
draft-ietf-webbotauth-httpsig-protocol-01, the adopted working-group draft that replaced the individual Meunier drafts, was posted on 2026-10-29. - 2026-10-27OpenAI updates the ChatGPT-User IP list
openai.com/chatgpt-user.json carried creationTime 2026-10-27 when read on 2026-11-04. - 2026-10-22IETF drafts propose a common IP-range format and crawler best practices
New webbotauth-related drafts include draft-illyes-webbotauth-jafar-01, a JSON format for publishing IP ranges of automated clients, and draft-illyes-webbotauth-cbcp-01 on crawler best practices, alongside the SAIL source address identity list draft. - 2026-10-21Anthropic's Claude Duet starts browsing with a Claude-Agent token
Claude Duet launched as a waitlist research preview for US Pro and Max subscribers. - 2026-10-20Anthropic refreshes its crawler IP list
claude.com/crawling/bots.json, the list Anthropic links for verifying ClaudeBot, Claude-User and Claude-SearchBot, carried creationTime 2026-10-20 when read, starting with 216.73.216.0/22. - 2026-10-15DataDome Q3 report: 25 billion AI agent requests
DataDome counted nearly 25 billion AI agent requests in Q3 2026, up 41.5 percent on the quarter, with Meta-ExternalAgent and Meta-WebIndexer at 49.7 percent. - 2026-10-01DataDome joins Experian Agent Trust
Experian announced DataDome joined its Agent Trust ecosystem, using verified Know Your Agent signals for continuous intent checks during a session. - 2026-09-17Arcjet launches Agent Runtime Security
Arcjet added runtime controls for AI agents that observe activity through OpenTelemetry, enforce policies on tool calls, APIs and databases, and keep an audit trail. - 2026-09-15Cloudflare adds Disallow AI Training and the Accountable label
Cloudflare's Training control gains Disallow AI Training, which writes a no-training preference into robots.txt and keeps Accountable mixed-use crawlers (Googlebot, Applebot, Bingbot) in search. - 2026-09-08Meta launches Muse, a personal agent that does not identify itself
Meta released Muse in the US on iOS, Android, the web and WhatsApp. - 2026-09-02HUMAN publishes Agent Trust Levels
HUMAN's AgenticTrust now grades observed agents High, Medium or Low, with High reserved for agents verifiable by a cryptographic key on each request. - 2026-08-28Cloudflare opens BotBase for Operators
Bot operators get a dashboard home for submissions under Application Security, with status (waiting, accepted, rejected with reasons), editing of existing entries, and a behaviour model to declare how a bot uses content. - 2026-08-19Google retires the Google-NotebookLM user agent
Google's user-triggered fetcher list now names Google-GeminiNotebook for URLs users add as sources and marks the former Google-NotebookLM string as supported until August 2026.
Sources
- IETF webbotauth working group documents
- vercel.com: block ai crawlers and assistants independently with managed rules
- openai.com: chatgpt user (JSON)
- verisoul.ai: claude duet
- claude.com: bots (JSON)
- helpnetsecurity.com: datadome ai agent traffic report
- experianplc.com: datadome joins experian agent trust ecosystem to add continuous
- arcjet.com: introducing agent runtime security observe enforce audit
- Have it both ways: stay discoverable in search while disallowing AI training
- verisoul.ai: meta muse
- humansecurity.com: evaluating ai agent trust humans agent trust levels
- BotBase for Operators
- Google user-triggered fetchers
Last reviewed 2026-11-04