--- title: "Shell: the 360 check | Agent Ready Checklist" description: "Shell on Ora's agent-readiness scan, read 2026-11-09: 52 of 100, grade C. 5 of 22 checklist rows met, 14 not met. One company a day." canonical: https://agentreadychecklist.com/blog/shell-agent-readiness site: "Agent Ready Checklist" last_updated: 2026-11-09 date: 2026-11-09 section: "360 checks" --- # Shell: the 360 check | Agent Ready Checklist > Shell on Ora's agent-readiness scan, read 2026-11-09: 52 of 100, grade C. 5 of 22 checklist rows met, 14 not met. One company a day. Source: https://agentreadychecklist.com/blog/shell-agent-readiness Site: Agent Ready Checklist Last reviewed: 2026-11-09 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). # Shell against the checklist: the 360 agent-readiness check 360 check, 9 November 2026. Shell is the London-based energy group, Europe's largest oil company by revenue. Shell scores 52 of 100 on Ora's agent-readiness scan, a C, read 9 November 2026. At the door, site accessible to 4 agent user-agents (ChatGPT-User, OAI-SearchBot, Claude-User, PerplexityBot). 5 of the 22 checklist rows with an Ora check are met. Twelve characters of home page, under a developer portal that search can find. On 9 November 2026 the public site at shell.com scored **52 of 100** on [Ora's](https://ora.ai/?utm_source=agentreadychecklist.com&utm_medium=referral&utm_campaign=check) agent-readiness scan, a C. Ora probes a site the way an agent would: it requests pages with the agents' own user agents, reads robots.txt, looks for the files agents read, and tries the API, MCP and payment surfaces. Below, each of the 31 checklist rows is read against the Ora checks the [crosswalk](https://agentreadychecklist.com/checklist/ora-checks/) names for it, so the result says which requirements this company meets, not only how many points it scored. Ora's one-line summary: Shell offers strong developer resource discoverability and brand presence, but lacks a published OpenAPI specification and MCP server for agent access. ## The score | Layer | Points | Share | What the layer asks | | --- | --- | --- | --- | | Discovery | 14 of 15 | 93% | Can agents find and recommend it | | Access | 26 of 57 | 46% | Can agents read it | | Usability | 17 of 48 | 35% | Can agents use it | | Payments | 0 of 0 | none apply | Can agents pay | | Total | 52 of 100 | 52% | Grade C | ## Row by row A row is met when every Ora check the crosswalk names for it passes, not met when every one fails, and partly when the checks disagree or one is partial. A row with no Ora check in the catalogue, or whose checks did not apply to this site, is marked as such; the [Verify Kit](https://agentreadychecklist.com/kit/) tests those by hand. | Row | Level | Result | Ora checks, and what Ora saw | | --- | --- | --- | --- | | AR-FIND-01; Do not block agents | MUST | Met | robots.txt AI crawler policy (pass), robots.txt agent-user policy (pass), Agent crawler reachability (pass), Not blocked by bot detection (pass).; AI search crawlers: allowed in 3 of 3 checks (3 crawlers on 1 page). AI training crawlers: allowed in 2 of 2 checks (2 crawlers on 1 page), not scored. | | AR-FIND-02; Answers on one first-party page | SHOULD | No Ora check | No check in Ora's catalogue names this row. | | AR-FIND-03; Sitemap listed in robots.txt | SHOULD | Met | Sitemap exists (pass).; Valid sitemap found at https://www.shell.com/.sitemap.xml with 2066 entries | | AR-FIND-04; AI Catalog at .well-known | MAY | Not met | AI Catalog discovery (not applicable), ARD discovery (fail).; No /.well-known/ard.json, no same-site rel="ard" link, and no legacy /.well-known/ai-catalog.json | | AR-READ-01; Answer in the initial HTML | MUST | Partly | Content without JavaScript (partial).; Homepage: The homepage has little text without JavaScript (12 chars), though nothing in it shows it needs JavaScript to render. | | AR-READ-02; Real HTTP status codes | MUST | Met | Redirect hygiene (pass).; Redirect hygiene: pass, 1 of 1. | | AR-READ-03; Home page with real links | SHOULD | Met | Public API/docs linked from homepage (pass).; Documentation site found at https://developer.shell.com | | AR-READ-04; Docs that answer agent tasks | SHOULD | No Ora check | No check in Ora's catalogue names this row. | | AR-READ-05; Fenced, language-tagged code | SHOULD | No Ora check | Code fence validity (not applicable). | | AR-READ-06; Link the discovery files | SHOULD | No Ora check | No check in Ora's catalogue names this row. | | AR-READ-07; Ship llms.txt | SHOULD | Not met | llms.txt exists (fail), llms.txt formatting (fail), llms.txt links resolve (not applicable).; No llms.txt found at /llms.txt or /.well-known/llms.txt. | | AR-READ-08; JSON-LD on key pages | SHOULD | Not met | JSON-LD structured data (fail).; No JSON-LD structured data found on homepage | | AR-READ-09; Advertise a markdown mirror | MAY | Not met | Markdown alternate link (fail).; No sampled page advertises a markdown alternate. Add (or the equivalent Link response header) pointing at a markdown twin of the page. | | AR-READ-11; Pricing as first-party HTML | SHOULD | Not met | Pricing info accessible (fail).; No pricing page or pricing data found | | AR-ACT-01; Auth an agent can finish | SHOULD | Partly | OAuth 2.0 support (pass), Scoped permissions (fail).; OpenID Connect discovery endpoint found at https://sso.shell.com | | AR-ACT-02; OpenAPI contract | SHOULD | Not met | OpenAPI spec published (fail).; No OpenAPI/Swagger specification found | | AR-ACT-03; MCP over Streamable HTTP | SHOULD | Not met | MCP server / manifest (fail), Listed in MCP registries (not applicable), MCP PKCE S256 support (not applicable), MCP OAuth metadata (not applicable).; No MCP server or manifest found | | AR-ACT-04; MCP server card | MAY | Not met | MCP server-card.json (fail).; No MCP server card found (checked /.well-known/mcp/server-card.json, the site's ai-catalog.json, servers[].card in mcp.json, and /server-card) | | AR-ACT-05; MCP Apps view | MAY | No Ora check | MCP Apps support (not applicable). | | AR-ACT-06; SDKs and a CLI | SHOULD | Partly | CLI tool available (fail), Multi-language SDK packages (partial), NPM/PyPI SDK package (pass).; No CLI tool found | | AR-ACT-07; WebMCP tools | SHOULD | Not met | WebMCP support (fail).; No WebMCP support detected - no document.modelContext / navigator.modelContext usage found (scanned 1 of 1 same-origin script bundle(s) for modelContext registrations) | | AR-ACT-08; Agent payments | MAY | No Ora check | x402 payment protocol (not applicable), MPP payment protocol (not applicable), UCP - Universal Commerce Protocol (not applicable), ACP - Agentic Commerce Protocol (not applicable), ACP delegate payment (not applicable). | | AR-ACT-09; NLWeb /ask endpoint | MAY | Not met | NLWeb /ask endpoint (fail).; No NLWeb-conformant /ask endpoint found via GET or POST | | AR-ACT-10; An MCP server or a CLI | MUST | Not met | MCP server / manifest (fail), CLI tool available (fail).; No MCP server or manifest found | | CF-CONTENT-SIGNALS; Content-Signal line | cloudflare | No Ora check | No check in Ora's catalogue names this row. | | CF-MARKDOWN-NEGOTIATION; Markdown by Accept | cloudflare | Not met | Markdown content negotiation (acceptmarkdown.com) (fail), Markdown agent docs (fail).; Homepage https://shell.com does not meet the Markdown negotiation requirements: Accept: text/markdown returned text/html;charset=utf-8; Vary header missing Accept (got "accept-encoding") | | CF-PAY-PER-CRAWL; Pay per crawl, HTTP 402 | cloudflare | No Ora check | No check in Ora's catalogue names this row. | | CF-OAUTH-DISCOVERY; OAuth discovery files | cloudflare | Not met | OAuth Protected Resource metadata (RFC 9728) (fail).; No /.well-known/oauth-protected-resource found (RFC 9728) | | CF-AGENT-SKILLS; Agent Skills index | cloudflare | No Ora check | Agent Skills index conformance (v0.2.0) (not applicable). | | VERCEL-NEGOTIATION-HEADERS; Markdown headers | vercel | Not met | Markdown content negotiation (acceptmarkdown.com) (fail).; Homepage https://shell.com does not meet the Markdown negotiation requirements: Accept: text/markdown returned text/html;charset=utf-8; Vary header missing Accept (got "accept-encoding") | | VERCEL-ROBOTS-EXCLUSIONS; Robots exclusions | vercel | Met | robots.txt AI crawler policy (pass), robots.txt agent-user policy (pass).; AI search crawlers: allowed in 3 of 3 checks (3 crawlers on 1 page). AI training crawlers: allowed in 2 of 2 checks (2 crawlers on 1 page), not scored. | ## The fixes worth most points Ora attaches an estimate to each failing or partial check: how many points the site would gain by passing it. These are the 5 with the largest estimates for Shell, each tied to the checklist row whose recipes cover it. MCP server / manifest Fail, 0 of 6, about 5.6 points to gain. Ora saw: No MCP server or manifest found. Row [AR-ACT-03](https://agentreadychecklist.com/checklist/ar-act-03/), MCP over Streamable HTTP. Scoped permissions Fail, 0 of 5, about 4.6 points to gain. Ora saw: No declared OAuth scopes, security schemes, or scoped-permission documentation found. Row [AR-ACT-01](https://agentreadychecklist.com/checklist/ar-act-01/), Auth an agent can finish. WebMCP support Fail, 0 of 5, about 4.6 points to gain. Ora saw: No WebMCP support detected - no document.modelContext / navigator.modelContext usage found (scanned 1 of 1 same-origin script bundle(s) for modelContext registrations). Row [AR-ACT-07](https://agentreadychecklist.com/checklist/ar-act-07/), WebMCP tools. OpenAPI spec published Fail, 0 of 7, about 4.1 points to gain. Ora saw: No OpenAPI/Swagger specification found. Row [AR-ACT-02](https://agentreadychecklist.com/checklist/ar-act-02/), OpenAPI contract. Public API with reachable endpoints Partial, 3 of 7, about 3.7 points to gain. Ora saw: API described in documentation at https://developer.shell.com but no machine-verifiable API surface confirmed. Best-of-protocols score: 3/7. No checklist row names this check. ## Our read Shell opens the door and gives an agent twelve characters to read. All four agent user agents reach shell.com, robots.txt allows the search and training crawlers and the sitemap at /.sitemap.xml lists 2,066 pages, but the home page renders twelve characters of text without JavaScript, the five sampled pages read without substantive content, and there is neither JSON-LD nor an llms.txt. The Discovery layer is nearly perfect at 14 of 15, which is what lifts the C. The act layer is a portal: developer.shell.com is real, OpenID Connect discovery sits at sso.shell.com and an SDK is on npm, though the API is described without a machine-verifiable surface and Ora found neither an OpenAPI specification nor an MCP server, and no declared scopes, about 14 points. The domain ranks fifth for Shell Global. Europe's largest oil company by revenue has a developer portal agents can find and a home page they cannot read; the template is the row to fix. ## Verify your own site The scan above is free and needs no account: enter a domain at [ora.ai](https://ora.ai/?utm_source=agentreadychecklist.com&utm_medium=referral&utm_campaign=check) and read the same four layers for your own site, then take each failing row to its recipe on this site. The command below runs the same audit from a terminal and exits with 1 under the score you set, so it can gate a deploy. Payments did not apply to this company; a site that sells is read on that layer too. [Why Ora, and its limits](https://agentreadychecklist.com/verify-with-ora/). ```bash npx ax@0.10 audit https://example.com --min-score 70 ``` [All 360 checks](https://agentreadychecklist.com/blog/), one company a day. Corrections to [the corrections page](https://agentreadychecklist.com/corrections/). ## Sources - [ora.ai/api/score/shell.com, Ora's scan (vendor)](https://ora.ai/api/score/shell.com?utm_source=agentreadychecklist.com&utm_medium=referral&utm_campaign=check)read 2026-11-09 - [Ora's scoring rules, version 2 (vendor)](https://ora.ai/api/ranker/releases/2/rules?utm_source=agentreadychecklist.com&utm_medium=referral&utm_campaign=check)read 2026-11-09 - [ora.ai/api/checks, contract 1.38.0 (vendor)](https://ora.ai/api/checks?utm_source=agentreadychecklist.com&utm_medium=referral&utm_campaign=check)read 2026-11-05 - [shell.com, the company's site](https://www.shell.com/)read 2026-11-09 Edition Data 2026-11-04 Reviewed 2026-11-05 Spec AgentReady v2.0, 273d2d7 Crawl 13,071 rows, 2026-11-04 Ora 132 checks, contract 1.38.0, 2026-11-05 Verify with Ora The scanner this site recommends for the whole checklist. 132 checks, a score out of 100, no account. `npx ax@0.10 audit --min-score ` [Why Ora, and its limits](https://agentreadychecklist.com/verify-with-ora/) Method - [How rows, commands and rates are made](https://agentreadychecklist.com/method/) - [How an Ora check is matched to a row](https://agentreadychecklist.com/checklist/ora-checks/) - [Pass rates by rank band](https://agentreadychecklist.com/data/by-rank/) Sources and read dates sit at the foot of the page. Last reviewed 2026-11-09. AgentReady v2.0, read 2026-11-04.