--- title: "Glossary | Agent Ready Checklist" description: "40 terms used on the checklist, from AR identifiers and conformance to llms.txt, MCP server cards and WebMCP, each with a sourced code example." canonical: https://agentreadychecklist.com/glossary site: "Agent Ready Checklist" last_updated: 2026-11-05 section: "Glossary" --- # Glossary | Agent Ready Checklist > 40 terms used on the checklist, from AR identifiers and conformance to llms.txt, MCP server cards and WebMCP, each with a sourced code example. Source: https://agentreadychecklist.com/glossary Site: Agent Ready Checklist Last reviewed: 2026-11-05 Agent endpoints: https://agentreadychecklist.com/llms.txt (index), https://agentreadychecklist.com/openapi.json (API), https://agentreadychecklist.com/mcp (MCP server), https://agentreadychecklist.com/ask (search) Access: https://agentreadychecklist.com/auth.md (no key). Pricing: https://agentreadychecklist.com/pricing.md (free). # Glossary 40 terms from the checklist, each defined in a sentence or two with a code example you can paste. The terms cover the spec vocabulary (MUST, SHOULD, MAY, conformance), the files agents read (robots.txt, llms.txt, JSON-LD, OpenAPI) and the protocols behind the Act rows (MCP, OAuth metadata, WebMCP). ## All 40 terms - [.well-known](https://agentreadychecklist.com/glossary/well-known/)URI prefix defined in RFC 8615 for site-wide metadata at fixed paths. Agents reached one in about 23% of traced runs. - [429 Too Many Requests](https://agentreadychecklist.com/glossary/429-too-many-requests/)Status code for throttling, defined in RFC 6585, optionally with Retry-After. AgentReady wants it in place of a challenge page. - [Agent Skills index](https://agentreadychecklist.com/glossary/agent-skills-index/)JSON at /.well-known/agent-skills/index.json listing skill documents with name, type, url and sha256 digest, so agents learn how to do tasks on a… - [AgentReady](https://agentreadychecklist.com/glossary/agentready/)Open, MIT-licensed spec of what a site or product implements so agents can find, read and act on it. Requirements carry AR-\* identifiers at MUST,… - [AI Catalog](https://agentreadychecklist.com/glossary/ai-catalog/)JSON file at /.well-known/ai-catalog.json listing every agent-facing resource a site ships. AR-FIND-04 (MAY). Uses the ARD data model. - [API catalog (RFC 9727)](https://agentreadychecklist.com/glossary/api-catalog/)A linkset document at /.well-known/api-catalog listing a publisher's APIs with links to their OpenAPI files and docs. - [AR identifier](https://agentreadychecklist.com/glossary/ar-identifier/)Stable requirement ID of the form AR--NN with FIND, READ or ACT. Never reused: a removed ID is retired, and new requirements take the next… - [Conformance](https://agentreadychecklist.com/glossary/conformance/)A site conforms to AgentReady when it meets every MUST that applies to it. AR-ACT applies only to products agents act on; an informational site… - [Content Signals](https://agentreadychecklist.com/glossary/content-signals/)A robots.txt directive declaring how content may be used after access: ai-train, search and ai-input, each yes or no. A stated preference, not an… - [Fenced code block](https://agentreadychecklist.com/glossary/fenced-code-block/)Code set apart from prose and tagged with its language, in markdown with backticks and in HTML with a language-\* class. AR-READ-05. - [Initial HTML](https://agentreadychecklist.com/glossary/initial-html/)The HTML the server returns before any script runs. AR-READ-01 requires the answer to be in it, server-rendered or prerendered. - [JSON-LD](https://agentreadychecklist.com/glossary/json-ld/)JSON format for linked data, embedded in a script tag with schema.org types. AR-READ-08 asks for it on key pages with facts repeated in visible text. - [Link header](https://agentreadychecklist.com/glossary/link-header/)HTTP response header (RFC 8288) carrying typed links, so an agent finds resources without parsing HTML. - [llms-full.txt](https://agentreadychecklist.com/glossary/llms-full-txt/)A companion file holding the full text of the docs in one markdown file. Large ones overflow agent context windows; Cloudflare splits per directory… - [llms.txt](https://agentreadychecklist.com/glossary/llms-txt/)A markdown file at /llms.txt: an H1 name, a blockquote summary and H2 sections listing key pages with descriptions. AR-READ-07 asks for a lean one. - [Markdown mirror](https://agentreadychecklist.com/glossary/markdown-mirror/)A .md twin of an HTML page, advertised with a link rel=alternate tag or Link header (AR-READ-09). - [Markdown negotiation](https://agentreadychecklist.com/glossary/markdown-negotiation/)Serving a markdown representation of a page when the request sends Accept: text/markdown, and HTML otherwise. Cloudflare checks it; AgentReady does… - [MAY](https://agentreadychecklist.com/glossary/may/)RFC 2119 keyword for an optional item. AgentReady uses it for emerging conventions with early, unmeasured adoption. - [MCP server](https://agentreadychecklist.com/glossary/mcp-server/)A server speaking the Model Context Protocol that exposes tools, resources and prompts to agent hosts. AgentReady wants one over Streamable HTTP,… - [MCP server card](https://agentreadychecklist.com/glossary/mcp-server-card/)JSON describing an MCP server before connection: name, version, transport endpoint, capabilities. Draft (SEP-2127); AR-ACT-04 (MAY). - [MUST](https://agentreadychecklist.com/glossary/must/)RFC 2119 keyword for an absolute requirement. In AgentReady v2.0 four requirements are MUST: the two changes that broke an agent, honest status… - [NLWeb /ask](https://agentreadychecklist.com/glossary/nlweb-ask/)An endpoint that takes a natural-language question and returns schema.org-shaped results, mirrored as an ask tool over MCP. AR-ACT-09 (MAY). - [OAuth authorization server metadata](https://agentreadychecklist.com/glossary/oauth-authorization-server-metadata/)RFC 8414 JSON at /.well-known/oauth-authorization-server naming the issuer and endpoints, so a client can start a flow without reading docs. - [OpenAPI](https://agentreadychecklist.com/glossary/openapi/)Machine-readable contract for an HTTP API: paths, parameters, responses, auth. AR-ACT-02 asks for one; agents found it by link 97% of the time. - [Pay per crawl](https://agentreadychecklist.com/glossary/pay-per-crawl/)Cloudflare AI Crawl Control feature (closed beta) that charges crawlers per request: 200 with payment intent, or 402 Payment Required with pricing. - [PKCE](https://agentreadychecklist.com/glossary/pkce/)Proof Key for Code Exchange (RFC 7636): the client sends a hashed code\_challenge, then proves it with the code\_verifier. Required by MCP… - [Plain-fetch client](https://agentreadychecklist.com/glossary/plain-fetch-client/)An HTTP client that downloads the response and does not run JavaScript, like most agent fetch tools. It sees only the initial HTML. - [Protected resource metadata](https://agentreadychecklist.com/glossary/protected-resource-metadata/)RFC 9728 JSON at /.well-known/oauth-protected-resource naming which authorization servers protect a resource. MCP servers must publish it. - [Retry-After](https://agentreadychecklist.com/glossary/retry-after/)Response header telling a client how many seconds (or until what date) to wait before retrying. Pair it with 429. - [robots.txt](https://agentreadychecklist.com/glossary/robots-txt/)Plain-text file at the site root that tells compliant crawlers which paths to skip, standardised as RFC 9309. Crawlers obey the most specific… - [SHOULD](https://agentreadychecklist.com/glossary/should/)RFC 2119 keyword for a recommendation you may skip only with a known reason. AgentReady uses it for practices that measurably shaped where agents… - [Sitemap](https://agentreadychecklist.com/glossary/sitemap/)XML file listing a site's canonical URLs for crawlers. One file holds up to 50,000 URLs; larger sites use a sitemap index. Reference it from… - [Soft 404](https://agentreadychecklist.com/glossary/soft-404/)A missing page answered with 200 and an error-looking body. Agents cannot tell it from a live page, so AR-READ-02 forbids it. - [Streamable HTTP](https://agentreadychecklist.com/glossary/streamable-http/)The MCP transport over a single HTTP endpoint that accepts JSON-RPC POSTs and may answer with JSON or a server-sent event stream. - [Training crawler](https://agentreadychecklist.com/glossary/training-crawler/)A crawler that collects pages to train models, such as GPTBot or ClaudeBot. Blocking it does not remove a site from AI search answers. - [User-agent group](https://agentreadychecklist.com/glossary/user-agent-group/)A block in robots.txt starting with one or more User-agent lines. A crawler uses only the most specific group that names it, ignoring the \* group. - [User-initiated fetcher](https://agentreadychecklist.com/glossary/user-initiated-fetcher/)A fetcher that loads a page because a user asked, such as ChatGPT-User or Claude-User. It may not consult robots.txt at all. - [Vary: Accept](https://agentreadychecklist.com/glossary/vary-accept/)Response header telling caches the body depends on the Accept request header, so HTML and markdown variants are stored apart. - [Web Bot Auth](https://agentreadychecklist.com/glossary/web-bot-auth/)IETF work in which a bot signs its HTTP requests and publishes public keys, so sites can verify which agent is calling. Keys live at a well-known… - [WebMCP](https://agentreadychecklist.com/glossary/webmcp/)Browser API (W3C Community Group draft) letting a page register tools an in-browser agent can call with the user session. AR-ACT-07 (SHOULD). ## Sources - [github.com/agentready-org/standard/blob/main/spec/spec.md](https://github.com/agentready-org/standard/blob/main/spec/spec.md)read 2026-11-04 - [llmstxt.org](https://llmstxt.org/)read 2026-11-04 Last reviewed 2026-11-05. AgentReady v2.0, read 2026-11-04.