· Nov 5, 2026
Cloudflare: Agents aren’t humans (but they kinda are)
Markdown · JSON · Original source
argues that remotely hosted AI Agents make it harder for Sites to know which User is acting, what that User is allowed to do, and who is responsible when something goes wrong. An Agent can act for several Users, while a Site sees the Agent or its Operator instead; blocking that intermediary can also block legitimate Users. Cloudflare illustrates the problem with a personal assistant whose errands mix clients’ preferences, purchases, and potentially abusive actions. The Site, it says, then bears costs involving fraud disputes, compliance, resource use, and content accessed without a visit or purchase. Cloudflare’s proposed principle is to preserve the relationship between Users and Sites even when an Agent sits between them. It argues that human checks and broad access permissions do not establish who the Agent acts for or who answers for its mistakes. Instead, it calls for interoperable protocols that can convey Agent identity, User information or reputation when needed, and Site rules in a machine-readable form, with varying levels of disclosure. Cloudflare presents the Trusted Agent Delegation coalition as a forum for developing this work, alongside standards efforts. It also describes Cloudflare products that it says are already in production, including tools to identify Agents and connect some requests to Users. The post says these are the first building blocks,
Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.
Original source
Blog · Article
Source details
Original source date: Nov 5, 2026 (2026-11-05T17:03:04.051Z).
Date basis: Date stated by the publisher.
Added to Escape Meridian: 2026-11-09T05:24:23.080Z. First observed: 2026-11-05T17:03:25.388Z. Retrieved: 2026-11-09T05:11:04.681Z.
Referenced by
Recorded citations and public file attachments. A reference alone does not establish agreement or influence.
No public references recorded here yet.
Limits
The post presents ’s own analysis, proposed standards direction, coalition, and product descriptions; the evidence does not independently establish how well the products work in practice. The post focuses primarily on Agents hosted somewhere other than the User’s device; local Agents and Agents running in the User’s own browser are outside its primary scope.
Open question
How will a Site know when several intermediaries sit between a User and the Site, and how will responsibility be established for an Agent’s actions?
Source-stated
Cloudflare argues that remotely hosted Agents can obscure which User is acting and leave Sites bearing costs when things go wrong.
This removes the ability for the Site to recognize, serve, and hold accountable the user or users behind the Agent
Source-stated
Cloudflare says blocking intermediaries becomes less viable as agentic traffic grows, leaving Sites to decide whether to trust an intermediary without direct information about the User.
Once blocking is no longer viable, Site Operators are left with the problem that intermediaries have always created.
Source-stated
Cloudflare argues that human checks and broad access permissions do not resolve the central question of who the Agent acts for or who answers when it gets something wrong.
It doesn't say who answers when the Agent gets it wrong, which is the question the Site actually needs answered.
Source-stated
Cloudflare calls for interoperable protocols to help Sites and Agents determine allowed actions and communicate preferences and policies.
The web needs tools and protocols that allow Sites and Agents to collaborate to determine what behaviors and actions are allowed
Source-stated
Cloudflare says it is announcing the Trusted Agent Delegation coalition to discuss the problem set and bring more defined problems to standards bodies as needed.
We are also announcing the Trusted Agent Delegation (TAD) coalition.
Source-stated
Cloudflare says its customers can use production tools to identify Agents and, for some sites, connect requests to the User an Agent acts for.
These are the first building blocks of the vision, already in production.
Source and contributor context
Cloudflare · Publisher
The publisher is identified here. No individual author has been inferred.
Source media
Media declared by the original publisher; availability checked separately from the editorial reading.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Cloudflare
Cloudflare provides networking, security and computing services that help run websites and applications.
Editorial provenance
Version 1. Publication ID: e0984d8c-272d-4c27-8507-c6e4b08e134e.
Prepared by gpt-6-luna; grounding review by gpt-6-luna.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-09T05:24:23.080Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: useful_perspective.
Observation: 9f565ba741bd70f3b6a2ce8c13be14c384cdb4c3e2b77cdaa9d333f28d08d550. Source text hash: 287e92ae833d299a3542a99eb4d2a1402d4aaf6e64fd40e1e18a31b1d4f74b71.
Revision and source-check history
Version 1 · 2026-11-09T05:24:23.080Z
Cloudflare: Agents aren’t humans (but they kinda are)
argues that remotely hosted AI Agents make it harder for Sites to know which User is acting, what that User is allowed to do, and who is responsible when something goes wrong. An Agent can act for several Users, while a Site sees the Agent or its Operator instead; blocking that intermediary can also block legitimate Users. Cloudflare illustrates the problem with a personal assistant whose errands mix clients’ preferences, purchases, and potentially abusive actions. The Site, it says, then bears costs involving fraud disputes, compliance, resource use, and content accessed without a visit or purchase. Cloudflare’s proposed principle is to preserve the relationship between Users and Sites even when an Agent sits between them. It argues that human checks and broad access permissions do not establish who the Agent acts for or who answers for its mistakes. Instead, it calls for interoperable protocols that can convey Agent identity, User information or reputation when needed, and Site rules in a machine-readable form, with varying levels of disclosure. Cloudflare presents the Trusted Agent Delegation coalition as a forum for developing this work, alongside standards efforts. It also describes Cloudflare products that it says are already in production, including tools to identify Agents and connect some requests to Users. The post says these are the first building blocks,
Original source
Blog · Article
Source details
Original source date: Nov 5, 2026 (2026-11-05T17:03:04.051Z).
Date basis: Date stated by the publisher.
Added to Escape Meridian: 2026-11-09T05:24:23.080Z. First observed: 2026-11-05T17:03:25.388Z. Retrieved: 2026-11-09T05:11:04.681Z.
Limits
The post presents ’s own analysis, proposed standards direction, coalition, and product descriptions; the evidence does not independently establish how well the products work in practice. The post focuses primarily on Agents hosted somewhere other than the User’s device; local Agents and Agents running in the User’s own browser are outside its primary scope.
Open question
How will a Site know when several intermediaries sit between a User and the Site, and how will responsibility be established for an Agent’s actions?
Source-stated
Cloudflare argues that remotely hosted Agents can obscure which User is acting and leave Sites bearing costs when things go wrong.
This removes the ability for the Site to recognize, serve, and hold accountable the user or users behind the Agent
Source-stated
Cloudflare says blocking intermediaries becomes less viable as agentic traffic grows, leaving Sites to decide whether to trust an intermediary without direct information about the User.
Once blocking is no longer viable, Site Operators are left with the problem that intermediaries have always created.
Source-stated
Cloudflare argues that human checks and broad access permissions do not resolve the central question of who the Agent acts for or who answers when it gets something wrong.
It doesn't say who answers when the Agent gets it wrong, which is the question the Site actually needs answered.
Source-stated
Cloudflare calls for interoperable protocols to help Sites and Agents determine allowed actions and communicate preferences and policies.
The web needs tools and protocols that allow Sites and Agents to collaborate to determine what behaviors and actions are allowed
Source-stated
Cloudflare says it is announcing the Trusted Agent Delegation coalition to discuss the problem set and bring more defined problems to standards bodies as needed.
We are also announcing the Trusted Agent Delegation (TAD) coalition.
Source-stated
Cloudflare says its customers can use production tools to identify Agents and, for some sites, connect requests to the User an Agent acts for.
These are the first building blocks of the vision, already in production.
Source and contributor context
Cloudflare · Publisher
The publisher is identified here. No individual author has been inferred.
Source media
Media declared by the original publisher; availability checked separately from the editorial reading.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Cloudflare
Cloudflare provides networking, security and computing services that help run websites and applications.
Editorial provenance
Version 1. Publication ID: e0984d8c-272d-4c27-8507-c6e4b08e134e.
Prepared by gpt-6-luna; grounding review by gpt-6-luna.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-09T05:24:23.080Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: useful_perspective.
Observation: 9f565ba741bd70f3b6a2ce8c13be14c384cdb4c3e2b77cdaa9d333f28d08d550. Source text hash: 287e92ae833d299a3542a99eb4d2a1402d4aaf6e64fd40e1e18a31b1d4f74b71.
uncertain · 2026-11-09T05:24:23.080Z · anthropic/claude-opus-5.5
Publication: e0984d8c-272d-4c27-8507-c6e4b08e134e
attribution — pass: The reading attributes the argument, the proposed principle, the TAD coalition announcement and the product claims to Cloudflare, which is the publisher of the post. Source assertions remain framed as Cloudflare's claims.
We are also announcing the Trusted Agent Delegation (TAD) coalition.
summary — uncertain: The summary is cut off mid-sentence and ends with 'The post says these are the first building blocks,'. A reader is left without the closing point that more will follow as standards mature. Smallest repair: complete or remove that sentence. 'Operator' is also used without a plain explanation; the source defines it as the party that hosts and maintains the Agent. Otherwise the summary preserves the Ezra analogy, the Intermediation Principle and the critique of human-in-the-loop checks and broad permission scopes.
These are the first building blocks of the vision, already in production. More will follow as the standards mature.
date — pass: The originalPublished value, 2026-11-05T17:03:04Z, comes from the publisher page and matches the feed entry. Nothing indicates an earlier first publication.
support — pass: Each claim is source-stated and has a matching excerpt. The statement that blocking an Agent blocks legitimate Users is supported by the Traffic commingling section. The question about multiple intermediaries follows from the source's own open issue. The limitations correctly note that the product claims are unverified and that local Agents are outside the post's primary scope.
We must still define a mechanism that supports multiple intermediaries
media — pass: The selected media is the verified, direct link to the original Cloudflare blog post. It is the most relevant way for readers to access the source.
Challenge or correct this reading
Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.
Cloudflare provides networking, security and computing services that help run websites and applications.
Explore CloudflareCloudflare provides networking, security and computing services that help run websites and applications.
Explore CloudflareCloudflare provides networking, security and computing services that help run websites and applications.
Explore CloudflareCloudflare provides networking, security and computing services that help run websites and applications.
Explore CloudflareCloudflare provides networking, security and computing services that help run websites and applications.
Explore CloudflareCloudflare provides networking, security and computing services that help run websites and applications.
Explore CloudflareCloudflare provides networking, security and computing services that help run websites and applications.
Explore Cloudflare