--- title: Anthropic defined the standards inside Agent Plugins. So... description: Vercel, AWS, Cursor, Microsoft, OpenAI, and Google released Agent Plugins 1.0.0, a package format meant to standardize where skills and MCP servers live so... canonical: https://daily.dev/posts/anthropic-defined-the-standards-inside-agent-plugins-so-why-isn-t-it-helping-govern-the-format--sjbkrjl67 twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Anthropic defined the standards inside Agent Plugins. So why isn’t it helping govern the format? | daily.dev og:description: Vercel, AWS, Cursor, Microsoft, OpenAI, and Google released Agent Plugins 1.0.0, a package format meant to standardize where skills and MCP servers live so... og:url: https://daily.dev/posts/anthropic-defined-the-standards-inside-agent-plugins-so-why-isn-t-it-helping-govern-the-format--sjbkrjl67 og:image: https://api.daily.dev/og/posts/SJBKrjl67.png og:image:alt: Anthropic defined the standards inside Agent Plugins. So why isn’t it helping govern the format? og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Anthropic defined the standards inside Agent Plugins. So why isn’t it helping govern the format? **[The New Stack](https://daily.dev/sources/newstack)** · 10 min read · 0 upvotes · 0 comments ## Summary Vercel, AWS, Cursor, Microsoft, OpenAI, and Google released Agent Plugins 1.0.0, a package format meant to standardize where skills and MCP servers live so they don't need to be forked per client. But the spec only governs file locations, not behavior: it defers skill format to Agent Skills and MCP wire behavior to the Model Context Protocol, both of which originated at Anthropic, which is notably absent from the Agent Plugins steering committee. Because conformance only requires supporting at least one component type (skills or MCP) and one transport, two conformant clients can share nothing in practice. Additional gaps include a manifest schema that contradicts the spec's own text on unknown fields, a client-extension namespace that can hide most of a plugin's real functionality outside the portable core, no portable secrets or credential mechanism, and unresolved PATH/environment behavior for stdio servers. A future v1.1 would need a conformance test suite, capability profiles, permissions, and signatures to close these gaps. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What does Agent Plugins 1.0 actually standardize for AI agent skills and MCP servers? Agent Plugins 1.0, released by Vercel with AWS, Cursor, Microsoft, and OpenAI (Google joined the same day), standardizes only file locations: a plugin.json manifest at the root, skills in a skills/ directory, and MCP servers in mcp.json. It does not define skill format (deferred to the Agent Skills spec) or MCP wire behavior (deferred to the Model Context Protocol), both of which originated at Anthropic. _Track the evolving Agent Plugins spec and its interoperability gaps on daily.dev before wiring plugins into production agents._ ### Why can two clients both claim Agent Plugins conformance but not actually run the same plugin? Because Section 11.1 of the Agent Plugins spec only requires a conformant client to support at least one component type, skills or MCP servers, and at least one transport (stdio or Streamable HTTP). A skills-only client and an MCP-only client can both be fully conformant while sharing no overlapping capability, so a plugin bundling both a skill and a stdio server may run in full on neither. _Developers picking agent clients can weigh these conformance gaps before committing to a plugin ecosystem, via daily.dev._ ### Does Agent Plugins provide a way to handle secrets or credentials for authenticated MCP servers? No, version 1 of Agent Plugins defines no portable credential mechanism. Section 7.2.1 states header values are visible package data rather than a secret mechanism, plugins must not embed credentials in headers or env variables, and there is no OAuth configuration; authorization discovery and credential storage are entirely client-managed, meaning setup differs per client even though packaging is portable. _Anyone wiring authenticated MCP servers into agent plugins can follow spec updates on daily.dev to avoid client-specific setup surprises._ ## Similar posts on daily.dev - [Agent Plugins Explained: Skills and MCP in One Folder](https://daily.dev/posts/agent-plugins-explained-skills-and-mcp-in-one-folder-oszuhbwpc) · Medium · 4 upvotes · 0 comments - [Agent Skills: Anthropic’s Next Bid to Define AI Standards](https://daily.dev/posts/agent-skills-anthropic-s-next-bid-to-define-ai-standards-qcfz4gydy) · The New Stack · 2 upvotes · 0 comments --- Tags: [#ai-agents](https://daily.dev/tags/ai-agents), [#mcp](https://daily.dev/tags/mcp), [#anthropic](https://daily.dev/tags/anthropic) [View this post on daily.dev](https://daily.dev/posts/anthropic-defined-the-standards-inside-agent-plugins-so-why-isn-t-it-helping-govern-the-format--sjbkrjl67) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Anthropic defined the standards inside Agent Plugins. So why isn’t it helping govern the format?","url":"https://daily.dev/posts/anthropic-defined-the-standards-inside-agent-plugins-so-why-isn-t-it-helping-govern-the-format--sjbkrjl67","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/anthropic-defined-the-standards-inside-agent-plugins-so-why-isn-t-it-helping-govern-the-format--sjbkrjl67"},"datePublished":"2026-08-17T19:30:47.358Z","dateModified":"2026-08-17T19:31:16.487Z","description":"Vercel, AWS, Cursor, Microsoft, OpenAI, and Google released Agent Plugins 1.0.0, a package format meant to standardize where skills and MCP servers live so...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/85d3cda7edf5d68339afb497a0efd249?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/85d3cda7edf5d68339afb497a0efd249?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"The New Stack","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"The New Stack","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/newstack","url":"https://daily.dev/sources/newstack"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/anthropic-defined-the-standards-inside-agent-plugins-so-why-isn-t-it-helping-govern-the-format--sjbkrjl67","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai-agents,mcp,anthropic","timeRequired":"PT10M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"The New Stack","item":"https://daily.dev/sources/newstack"},{"@type":"ListItem","position":3,"name":"Anthropic defined the standards inside Agent Plugins. So why isn’t it helping govern the format?"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/anthropic-defined-the-standards-inside-agent-plugins-so-why-isn-t-it-helping-govern-the-format--sjbkrjl67#faq","mainEntity":[{"@type":"Question","name":"What does Agent Plugins 1.0 actually standardize for AI agent skills and MCP servers?","acceptedAnswer":{"@type":"Answer","text":"Agent Plugins 1.0, released by Vercel with AWS, Cursor, Microsoft, and OpenAI (Google joined the same day), standardizes only file locations: a plugin.json manifest at the root, skills in a skills/ directory, and MCP servers in mcp.json. It does not define skill format (deferred to the Agent Skills spec) or MCP wire behavior (deferred to the Model Context Protocol), both of which originated at Anthropic. Track the evolving Agent Plugins spec and its interoperability gaps on daily.dev before wiring plugins into production agents."}},{"@type":"Question","name":"Why can two clients both claim Agent Plugins conformance but not actually run the same plugin?","acceptedAnswer":{"@type":"Answer","text":"Because Section 11.1 of the Agent Plugins spec only requires a conformant client to support at least one component type, skills or MCP servers, and at least one transport (stdio or Streamable HTTP). A skills-only client and an MCP-only client can both be fully conformant while sharing no overlapping capability, so a plugin bundling both a skill and a stdio server may run in full on neither. Developers picking agent clients can weigh these conformance gaps before committing to a plugin ecosystem, via daily.dev."}},{"@type":"Question","name":"Does Agent Plugins provide a way to handle secrets or credentials for authenticated MCP servers?","acceptedAnswer":{"@type":"Answer","text":"No, version 1 of Agent Plugins defines no portable credential mechanism. Section 7.2.1 states header values are visible package data rather than a secret mechanism, plugins must not embed credentials in headers or env variables, and there is no OAuth configuration; authorization discovery and credential storage are entirely client-managed, meaning setup differs per client even though packaging is portable. Anyone wiring authenticated MCP servers into agent plugins can follow spec updates on daily.dev to avoid client-specific setup surprises."}}]} ```