Web Bot Auth
web bot auth
A proposed web standard, backed by Cloudflare, that lets an automated client prove which bot it is by signing its requests, so sites can let trusted agents in and keep others out.
In our briefings
The claim each of the newest briefings on Web Bot Auth would have you leave with.
An agent's identifier is the HTTPS URL where it publishes its keys, sent in the Signature-Agent header. The URL becomes an identifier once the verifier resolves it and finds a key that verifies the request; until then, the draft says, "verifiers MUST NOT attach policy to it." A valid signature then shows that a holder of a key the URL publishes signed the message, and the draft says it "says nothing about who operates the Agent, whether the Agent is benign, or whether the request is authorized," leaving those to the origin's policy. Nothing stops an agent from dropping a URL and standing up another. Origins can log, rate limit, allowlist or block the URL the way they treat IP addresses and User-Agent strings today.
Web Bot Auth's working-group draft names a bot by its key URL and defines no revocationBI-FIND-2026-1784
The draft's OAuth profile says an agent SHOULD use a Client ID Metadata Document as its client_id, and the metadata SHOULD carry a new agent_uri parameter naming the agent. It then warns that the parameter proves nothing by itself: "anyone can publish client metadata that names someone else's agent URI and then authenticate correctly as that client." An authorization server therefore MUST NOT treat a caller as an authenticated agent unless one of two routes succeeds in the current exchange. On the first, the agent's own document links to that exact client_id and the client authenticates with private_key_jwt using qualifying keys. On the second, the request carries a signature for the agent URI verified under the draft's request profile, and the agent attributed is the one that signature names; if the metadata's agent_uri names a different agent, the server MUST NOT display or record it. A consent screen can show a verified agent only if that check ran on a back-channel request before consent, and the draft says agents that want this SHOULD use Pushed Authorization Requests. Delegation SHOULD go through token exchange with the agent URI as sub in the act claim, and the agent's assertion keys MUST NOT double as DPoP keys.
Agent URI draft -04 ties an agent's name to a key document and bars OAuth servers from trusting the name aloneBI-FIND-2026-1809
In Bu's review method, each security claim an agent protocol makes, such as who authorized the task, which running instance is acting, what scope was delegated, whether a human approved and what evidence the action left, should get its own row naming what carries it, who verifies it, what it is bound to, how fresh it must be and what happens when it fails. Claims cannot be combined into a stronger one than their parts, and "a receipt, log entry, or audit record proves only the statement it records". A protocol that cannot fill a row can instead mark the claim as inherited, out of scope or left to deployment policy.
Bu's checklist names a verifier for each agent-protocol claim and checks 'the same agent as before' separatelyBI-FIND-2026-0860
›Cited in 8 briefings
Every briefing that discusses this term, newest first.
- Web Bot Auth's working-group draft names a bot by its key URL and defines no revocation
Draft · · IETF Datatracker · Thibault Meunier et al. · Oct 29, 2026
- Agent URI draft -04 ties an agent's name to a key document and bars OAuth servers from trusting the name alone
Draft · · IETF Datatracker · Yaswanth Narvaneni et al. · Oct 20, 2026
- Bu's checklist names a verifier for each agent-protocol claim and checks 'the same agent as before' separately
Draft · · IETF Datatracker · Songbo Bu · Oct 14, 2026
- World's maker proposes a site-scoped handle for each verified human, and bars using it alone to log anyone in
Draft · · IETF Datatracker · Aurélien Brézun, Tools for Humanity · Aug 6, 2026
- WorkOS's Tobin South, in November 2025: agents will get a new kind of account, not a new identity standard
Podcast · · CodingCat.dev Podcast · 5.10 - How to Secure AI - Identity Access Management for Agents · Nov 21, 2025
- OAuth handles an agent in one company, working while you watch. The OpenID paper says that is where it stops.
Article · · OpenID Foundation · Tobin South et al. · Oct 2025
- An OpenID Foundation co-chair says AI agents need modest upgrades to today's identity systems, not a new one
Podcast · · Identity at the Center · #390 - Identity Management for Agentic AI with Tobin South · Dec 8, 2025
- Stytch sells a one-call check for whether a visitor is an AI agent, priced at half a cent a lookup
Podcast · · So What About AI Agents · EP 39 - Agentic Authentication | Bobbie Chen · Sep 23, 2025
Related terms
What appears alongside Web Bot Auth in the same briefings.
BrowserbaseCloudflareTrusted Agent ProtocolDigest Fields (RFC 9530)Agentforce