---
type: "post"
id: "poppy-publishes-final-1-0-specification-after-sierra-381dc6"
slug: "poppy-publishes-final-1-0-specification-after-sierra-381dc6"
title: "Poppy publishes final 1.0 specification after Sierra’s announcement"
description: "Poppy’s final 1.0 specification, published November 14, 2026, turns the Personal Agent Protocol into a versioned implementation target and documents changes from its 0.4 release candidate."
retrieval_nugget: "Poppy’s final 1.0 specification, published November 14, 2026, turns the Personal Agent Protocol into a versioned implementation target and documents changes from its 0.4 release candidate. Its OAuth, consent, key-binding and compatibility rules matter to teams building agents that act across company services: implementers may need to change code, and companies must make authorization behavior safer and more predictable."
published_at: "2026-11-17"
updated_at: "2026-11-17"
record_date: "2026-11-17"
date_kind: "published_at"
topics: ["agent-protocols","agents"]
entities: ["Sierra","Poppy","Personal Agent Protocol","Shopify","Muse","Cloudflare Workers"]
related_patterns: ["agent-protocols-become-interoperability-layer"]
novelty_label: "new_fact"
what_changed: "The final Poppy 1.0 specification is now published, with changes including required conversation descriptions and optional pushed authorization request requirements."
prior_refs: ["https://newruntime.com/posts/sierra-releases-personal-agent-protocol-poppy-1-0-dd15ec/"]
source_url: "https://personalagentprotocol.org/docs/changelog"
source_urls: ["https://personalagentprotocol.org/docs/changelog"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "post:poppy-publishes-final-1-0-specification-after-sierra-381dc6"
status: "published"
visuals: []
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/posts/poppy-publishes-final-1-0-specification-after-sierra-381dc6/","markdown":"https://newruntime.com/posts/poppy-publishes-final-1-0-specification-after-sierra-381dc6.md","json":"https://newruntime.com/posts/poppy-publishes-final-1-0-specification-after-sierra-381dc6.json"}
markdown_url: "https://newruntime.com/posts/poppy-publishes-final-1-0-specification-after-sierra-381dc6.md"
---

# Poppy publishes final 1.0 specification after Sierra’s announcement

## Retrieval answer

Poppy’s final 1.0 specification, published November 14, 2026, turns the Personal Agent Protocol into a versioned implementation target and documents changes from its 0.4 release candidate. Its OAuth, consent, key-binding and compatibility rules matter to teams building agents that act across company services: implementers may need to change code, and companies must make authorization behavior safer and more predictable.

## Summary
Poppy’s final 1.0 specification, published November 14, 2026, turns the Personal Agent Protocol into a versioned implementation target and documents changes from its 0.4 release candidate. Its OAuth, consent, key-binding and compatibility rules matter to teams building agents that act across company services: implementers may need to change code, and companies must make authorization behavior safer and more predictable.

## What happened
The Poppy changelog says version 1.0 was published on November 14, 2026, and describes it as version 0.4 with fixes implementers needed. The specification marks some changes as breaking: an implementation that followed 0.4 may need to change even when a change is categorized as minor. The changelog also says later 1.x versions are intended to add features without breaking 1.0 implementations. From 1.0 onward, each side is to send only fields, values or events defined by the other side’s declared version, allowing companies to offer later-version behavior selectively.

Several changes govern sign-in and authorization. A company may require Pushed Authorization Requests (PAR) for Direct Sign-In, using `require_pushed_authorization_requests` in issuer metadata. In that case, the Personal Agent must push the authorization request with its client assertion and open the authorization endpoint with only `client_id` and `request_uri`. The agent should use PAR whenever the company offers it, and the pushed request includes a DPoP proof binding the code to a key. Since a `request_uri` expires quickly, the changelog recommends that a link opened later lead to the agent’s own page, which pushes the request when the user opens it.

The final rules also make consent more explicit. A company’s consent page must list every scope it will grant during Direct or Device Sign-In, and may mark previously granted scopes. A Direct Sign-In request must include `prompt=consent`, except when the redirect is received only on the device that opened the authorization URL. The company should display the consent page for Device Sign-In and when `prompt=consent` is used, even if the account previously granted every requested scope to that client ID. The stated concern is that every user of a Personal Agent shares one client ID: without a visible list of scopes, one user could approve a sign-in initiated by another without seeing the scopes being granted.

Poppy 1.0 also requires companies to accept client assertions and delegations signed with PS384 as well as ES256. For refresh tokens, a Personal Agent must prove possession of the same DPoP key used to obtain the token; it keeps that key unless the user signs in again or a required extension defines replacement. The changelog says a company may bind refresh tokens to that key and reject refreshes using another key. Other implementation details include a maximum five-minute lifetime for client assertions, a requirement for UCP scope requests to name every scope still needed, and a rule that `context: null` clears previously shared context.

The changelog describes compatibility and interaction behavior too. Conversation endpoints now require descriptions of what each protocol endpoint can help with, enabling an agent to choose between, for example, a company API and another provider’s support agent. Unknown conversation roles and event names have defined handling so later versions can extend the protocol. The meaning of `working` is narrowed to a company still replying; work that finishes later should be described as underway and followed by another message when complete. The changelog additionally says Mediated Sign-In is planned for retirement on March 17, 2027.

## What the archive adds
New Runtime covered the Poppy 1.0 release on November 14, noting Shopify’s enablement for all merchants, Muse’s prioritization and participation by more than 100 companies as design partners. This changelog makes the release more actionable by detailing implementation changes and migration risks: it is a new fact on a known story. [New Runtime’s earlier coverage](https://newruntime.com/posts/sierra-releases-personal-agent-protocol-poppy-1-0-dd15ec/).

## Trend signal
Poppy’s version negotiation and endpoint descriptions are consistent with the shift toward protocols as an interoperability layer: agents need to discover compatible capabilities and select among company services without assuming every endpoint is interchangeable. The changelog does not establish adoption or prove that integrations work reliably at scale, so this is a specification-level signal rather than evidence of mature deployment. [Agent protocols become the interoperability layer](https://newruntime.com/patterns/agent-protocols-become-interoperability-layer/).

## Why now
The changelog dates 1.0 to November 14, after the 0.4 release candidate published November 10. The final version therefore gives teams implementing the candidate a short interval to review changes before treating the protocol as a stable 1.0 target. It also sets a specific future milestone: the planned retirement of Mediated Sign-In on March 17, 2027.

## Connections to other events
New Runtime’s October 24 note described a GitHub demo running Poppy on Cloudflare Workers, an example of the protocol being implemented in a concrete environment. The 1.0 changelog now supplies the finalized rules that such implementations must target. [The earlier demo note](https://newruntime.com/posts/poppy-on-cloudflare-workers-github-repository-irvinebroque-poppy-c78edd/).

## Why it matters
A protocol can make agent-to-company integrations easier to compose, but only if the boundaries around identity, permission and compatibility are explicit. Poppy 1.0’s consent rules address a shared-client problem by requiring users to see every scope granted, including previously approved scopes. Its DPoP refresh requirement ties continued access to the key that received the refresh token, rather than allowing a copied token to be refreshed with a different key. Version declarations and defined handling of unknown events give implementers a path to add behavior while limiting surprises for older counterparts.

These are operational requirements, not just document cleanup. A team upgrading from 0.4 should audit sign-in flows, consent screens, supported signature algorithms, token storage and refresh logic, as well as any extension or conversation behavior it relies on. Companies exposing services should review whether scope descriptions explain what access enables and whether a task can work without it. The result is a more concrete integration contract, but teams still need to test that both sides implement it correctly.

## Limits and uncertainty
The changelog documents specification requirements and guidance; it does not establish how many companies or Personal Agents have implemented the final version, or whether implementations interoperate successfully in production. It identifies breaking changes for 0.4 implementers but does not quantify upgrade effort or compatibility failures. The security benefits of consent, PAR and DPoP depend on correct implementation by both sides, and the changelog is not an independent security audit. The stated Mediated Sign-In retirement date is a protocol plan, not evidence that every provider has migrated.

## Skill to build now
If your agent or company integration targets Poppy 0.4, make a migration checklist against 1.0 this week. Start with the breaking authorization changes: verify that consent lists all granted scopes, that refresh uses the original DPoP key, and that client assertions expire within five minutes. Then test PS384 acceptance and version negotiation in a staging environment before changing production traffic.

## Reader takeaway
Poppy 1.0 is more than a stable label: it fixes concrete rules for authorization, consent, token binding and version compatibility. Teams already implementing the protocol should treat the changelog as a migration document, not assume their 0.4 behavior is compatible by default.

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- topic hub: [Agent Protocols](https://newruntime.com/topics/agent-protocols.md): Explore the agent-protocols topic hub.
- topic hub: [Agents](https://newruntime.com/topics/agents.md): Explore the agents topic hub.
- collection: [Field Notes](https://newruntime.com/posts.json): Open the latest editorial analysis.
- collection: [Sources](https://newruntime.com/sources/): See the original sources this archive cites.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)
