HonestClaw
What is Agent Readiness?

Agent Readiness Score

openai.com

50 — Developing. Agents can find parts of this site, but key capabilities are hard to reach.

23 of 45 checks conclusive · Confidence Medium · Evidence coverage 51%

HonestClaw measured how well openai.com's site and APIs work for AI agents — across 9 dimensions and 45 automated checks.

Dimensions

Discoverability5/13
Content Accessibility3.5/5
Bot Access ControlNot yet scored
Protocol & Capability Discovery15/26
Authentication & Credentials10.3/19
Documentation Quality8.5/10
Structured Data I/O2.8/13
Observable State & Reliability5.3/15
Agent Safety & TrustNot yet scored

Checks

Discoverability

API existenceapi-existenceNo API endpoint found at common paths (only HTML pages or 404s).Not detected
API catalogapi-catalogNo RFC 9727 API Catalog found at /.well-known/api-catalog.Not detected
API versioningapi-versioningThe API version lifecycle is documented at https://developers.openai.com/api/docs/deprecations, including deprecation/sunset. (source: docs)Cross-verified
robots.txtrobots-txtNo live subdomains matched this check.Couldn't verify
sitemap.xmlsitemap-xmlNo live subdomains matched this check.Couldn't verify
Canonical domaincanonical-domainNo live subdomains matched this check.Couldn't verify
+ 1 more in the interactive report →

Content Accessibility

llms.txtllms-txtA curated llms.txt index is published on a non-top subdomain at https://developers.openai.com/llms.txt: 38 links across 4 sections. Its credit is capped at useful because it is not published on the main domain.Useful
llms-full.txtllms-full-txtNo live subdomains matched this check.Couldn't verify
Markdown twinmarkdown-twinNo live subdomains matched this check.Couldn't verify
ai.txtai-txtNo live subdomains matched this check.Couldn't verify
+ 3 more in the interactive report →

Bot Access Control

AI bot rulesai-bot-rulesNo live subdomains matched this check.Couldn't verify
Content signalscontent-signalsNo live subdomains matched this check.Couldn't verify
TDM reservationtdmrepNo live subdomains matched this check.Couldn't verify
+ 1 more in the interactive report →

Protocol & Capability Discovery

MCP supportmcp-supportFirst-party documentation at https://developers.openai.com/learn/docs-mcp explicitly states that the company operates MCP servers and explains both their capabilities and how clients connect or authenticate.Cross-verified
A2A Agent Carda2a-agent-cardNo A2A Agent Card found at the standard well-known locations.Not detected
MCP descriptormcp-descriptorInformational — not scoredNo MCP descriptor found at the standard locations an agent would check.Not detected
MCP Server Cardmcp-server-cardNo MCP Server Card found at the standard locations an agent would check.Not detected
OpenAPI specificationopenapi-specNo OpenAPI/Swagger specification found.Not detected
SDK availabilitysdk-availabilitySDK libraries page at https://developers.openai.com/blog/skills-agents-sdk: 4 GitHub repos, 0 registry links, 0 install commands (4 pointers).Cross-verified
Rate limit documentationrate-limit-docsRate limits are fully documented at https://developers.openai.com/api/docs/guides/rate-limits, with concrete limits, response headers, 429 handling, Retry-After, and tier/plan differences.Cross-verified
WebMCPwebmcpInformational — not scoredWebMCP browser probing is not yet available in this scanner release.Couldn't verify
Link headerslink-headersNo live subdomains matched this check.Couldn't verify
.well-known indexwell-known-indexNo live subdomains matched this check.Couldn't verify
+ 3 more in the interactive report →

Authentication & Credentials

Programmatic auth flowprogrammatic-auth-flowA programmatic auth flow is documented at https://developers.openai.com/workspace-agents/authentication, naming a non-interactive grant and a token endpoint. To reach the top level, it would need scopes and token expiry/refresh documented.Useful
Auth documentationauth-documentationAuthentication is fully documented at https://developers.openai.com/ads/api-reference/authentication, naming complete method(s) and how they are sent, corroborated across surfaces or covering multiple methods.Cross-verified
Credential managementcredential-managementNo credential management documentation found at the standard docs locations an agent would check.Not detected
Sandbox environmentsandbox-environmentA sandbox/testing page exists at https://learn.chatgpt.com/docs/windows/windows-sandbox, but it doesn't show how to obtain test credentials or a test base URL.Parseable

Documentation Quality

API reference depthapi-reference-depthThe API reference at https://developers.openai.com/api/reference/overview documents parameters with example requests and responses. To reach the top level, it would need request and response bodies, authentication, pagination, and error codes.Useful
Changelog presencechangelog-presenceA changelog is published at https://learn.chatgpt.com/docs/changelog with multiple recent dated entries, indicating it is actively maintained.Cross-verified

Structured Data I/O

Structured data I/Ostructured-data-ioNo published structured-data contract found (no typed spec response schemas).Not detected
Machine-readable pricingmachine-readable-pricingNo machine-readable pricing endpoint found (HTML pricing pages do not qualify).Not detected
Data export APIdata-export-apiA data export API is documented at https://developers.openai.com/ads/bulk-api with a bulk/export endpoint or export format. To reach the top level, it would need a bulk endpoint, async-job or pagination, and a stated export format.Useful

Observable State & Reliability

Status & health endpointsstatus-and-healthNo health or status endpoint found.Not detected
Retry & idempotencyretry-and-idempotencyNo retry or idempotency documentation found at the standard docs locations an agent would check.Not detected
Webhook documentationwebhook-documentationA webhook documentation page exists at https://developers.openai.com/api/docs/guides/webhooks, but it doesn't list specific event types with their payload shape.Parseable
Error documentationerror-documentationErrors are documented at https://developers.openai.com/plugins/deploy/submission-errors with specific codes or a documented response shape. To reach the top level, it would need recovery guidance and a machine-readable error format, or corroboration against the API spec.Useful
Machine payments (x402/MPP)x402-or-mpp-supportInformational — not scoredNo machine-payment signal observed (no HTTP 402, no payment-required response header, and no /.well-known/x402.json manifest). Payment support cannot be ruled out from an unauthenticated probe.Couldn't verify

Agent Safety & Trust

security.txtsecurity-txtNo live subdomains matched this check.Couldn't verify
Legal policy pageslegal-pagesNo live subdomains matched this check.Couldn't verify
Org identity signalsorg-identityNo live subdomains matched this check.Couldn't verify
+ 1 more in the interactive report →