§7 · Method · Spec status checked September 2026
How we research, compare and score agent readiness tools
Sources, scoring rules, and limitations, in one place.
Agent Readiness Compare editors · Spec status checked September 2026
Answer
This reference is desk research from public specification pages and vendor material, last reviewed September 2026. We compare tools by the job they do; we score only like-for-like tools, currently two readiness scanners, with six published criteria.
On this page
1.What does this reference cover?
Eight standards that affect whether an AI agent can discover, read, use, pay or identify itself to a website: robots.txt (RFC 9309) as applied to AI crawlers, llms.txt, agents.json, WebMCP, MCP, A2A, x402 and Web Bot Auth. We also cover the tools whose public pages document support for these standards: ora.ai, Cloudflare, Vercel, webmcp.com and nekuda, with shorter mentions of Forter and is-agentic.com. We left out vendors whose public pages we could not read or that did not describe agent readiness features at review time.
2.Where do the facts come from?
Every fact about a standard comes from its specification or maintainer page. Every fact about a tool comes from the vendor's own public pages or documentation. Each fact block ends with a source line and the review month. We did not interview vendors, and we did not use private material, customer data or sales calls as a source for anything published here.
Standards move quickly. Several covered here are drafts or community proposals (WebMCP, agents.json, Web Bot Auth). Each standard page states the status we found and the date.
3.How do we compare tools?
Most tools in this space do different jobs. A scanner that scores a site, a CDN feature that serves markdown, and an SDK that builds WebMCP tools are not substitutes, so ranking them in one list would mislead. On the Tools page we compare tools by job, using six status labels (section 5). We score tools only when they do the same job.
4.How do we score readiness scanners?
Two tools in our set are like-for-like: ora.ai Scan and Cloudflare's Is It Agent Ready. Both take a public URL and report which agent readiness signals a site exposes. is-agentic.com also does this, but its scans are run by ora.ai, so we describe it with ora.ai and do not score it separately.
Scores are an editorial assessment on a 0-5 scale per criterion, based on what each vendor's public pages document. The weighted total is computed from the published weights below. We did not run scans to produce these scores.
| Criterion | Weight | What we look for |
|---|---|---|
| Standards coverage | 20% | which discovery, content, capability, identity and commerce standards the scanner checks. |
| Real-agent task testing | 20% | does the tool run an agent through a task and show where it stopped, beyond checking for files and headers? |
| Published methodology | 15% | are the layers, check counts, weights and grade bands public? |
| Fix path on the same platform | 15% | can the findings be fixed with features the same vendor provides? |
| General availability | 10% | can anyone use the full tool today without a waitlist or sign-up? |
| Developer workflow | 20% | CLI, API or MCP access so a coding agent or CI job can run checks. |
Scanner score (editorial assessment, 0-5). Compares readiness scanners on what their public pages document. It is not a measure of overall product quality.
| Criterion | ora.ai | Cloudflare Is It Agent Ready |
|---|---|---|
| Standards coverage (20%) | 4.5 / 5AgentReady spec covers robots.txt, llms.txt, sitemaps, JSON-LD, markdown negotiation, MCP, OpenAPI, WebMCP, NLWeb, OAuth 2.0, Web Bot Auth, x402, ACP, UCP and MPP. | 4.5 / 5Checks robots.txt, sitemaps, DNS-AID, markdown negotiation, AI bot rules, Web Bot Auth, MCP, OAuth, Agent Skills, WebMCP, x402, MPP, UCP and ACP. |
| Real-agent task testing (20%) | 4.5 / 5Journey records a real agent attempting a chosen intent and shows where it stalled; Deep Scan v2 scores what agents did. | 2.0 / 5Checks published standards; the public page does not describe running an agent through a task. |
| Published methodology (15%) | 4.5 / 5Methodology publishes four layers, points per layer (20/30/40/10), check counts (16/41/56/6) and grade bands A+ to F. | 3.0 / 5Lists five categories and the standards checked; score bands are not published on the page. |
| Fix path on the same platform (15%) | 3.5 / 5Fixes are ranked by impact and the MCP server can rescan after each fix; implementation happens on your own stack. | 4.5 / 5Markdown for Agents, AI Crawl Control, Pay per crawl and Web Bot Auth verification are Cloudflare features on the same network. |
| General availability (10%) | 3.0 / 5Free tier needs no key or sign-up; the home page asks new users to join a waitlist. | 5.0 / 5Public scan page; enter a URL and scan. |
| Developer workflow (20%) | 4.5 / 5npx ax audit CLI, REST API, MCP server and a local WebMCP audit (npx ax webmcp-audit). | 2.0 / 5No CLI or API for the scanner is documented on the public page. |
| Weighted total | 4.2 / 5 | 3.3 / 5 |
ora.ai leads on: Real-agent task testing, Published methodology, Developer workflow. Cloudflare leads on: Fix path on the same platform, General availability. Tie: Standards coverage.
Source: ora.ai methodology · ora.ai llms.txt · ora.ai AgentReady post · journey.ora.ai · webmcp.ora.ai · isitagentready.com · Cloudflare Markdown for Agents · Cloudflare AI Crawl Control · Cloudflare Web Bot Auth · Reviewed Sep 2026
ora.ai scores lower than Cloudflare on two criteria: fix path on the same platform, and general availability. The two tie on standards coverage.
5.What do the status labels mean?
- Provides
- the tool's public pages document that it implements or serves this job.
- Checks
- the tool tests for this signal but does not implement it for you.
- Partial
- documented, with a stated limit (beta, a subset, or for crawlers only).
- Partner
- provided through a partnership with another listed vendor.
- Pending
- the vendor has a feature that is not yet documented on its public pages; to be completed before launch.
- Not documented
- we did not find it on the vendor's public pages at review time. It may exist.
6.What are the limitations?
Limitation
- Public sources only. No vendor interviews, no private documentation.
- No hands-on testing. We did not scan sites, run agents or measure outcomes.
- Coverage is a snapshot. Specifications and products changed often in 2026; each page states its review month.
- "Not documented" means we did not find it, not that it does not exist.
- Scores cover two like-for-like scanners only and are not a measure of overall product quality.
7.How do I report a correction?
Email editor@agentreadinesscompare.com with the page, the statement, and a public source. We review corrections within five business days and note material changes in the page's review line.
Frequently asked questions
Did you test these tools?
No. This is desk research from public vendor pages and specifications. We did not run scans or agents.
Why score only two tools?
Scores are meaningful only between tools that do the same job. ora.ai Scan and Cloudflare Is It Agent Ready both scan a public URL for readiness signals. The other tools serve, build or verify, so we compare them by job instead.