Shellshell.com
Shell's API Design and Agent Usability are strong foundations, but two areas are holding back the partner program: partners and their AI agents cannot discover or self-serve the API without a sales conversation, and the documentation surface lacks the machine-readable signals modern answer engines and coding agents depend on. Prioritize opening a self-serve onboarding path with working code examples first, then layer in the discovery and structured-data signals that let agents find and cite your APIs without human intervention.
API DesignA clean, typed, well-governed API contract agents can reason about2 pass1 warn0 fail83A
| Signal | Points | Findings | Rationale | |
|---|---|---|---|---|
| warn | Schema coverage & depthvia sdk | 15/25 | 0/4 (0%) public params are fully typed (no any/**kwargs). Analyzer reports not all methods have a derivable input schema. Investigated: sdk 60%. | Typed, complete request/response schemas are what make agent function-calling possible. |
| pass | Security & governance hygienevia sdk | 15/15 | Code is formatter-clean with zero linter errors and zero style warnings. Investigated: sdk 100%, wellknown 0%. | No leaked secrets, no critical lint violations, no OWASP API Top-10 spec smells, and a published vulnerability-disclosure channel. |
| pass | Example coveragevia sdk | 10/10 | README is present and includes a runnable quickstart. Investigated: sdk 100%, docs 0%. | Examples carry shape semantics schemas under-specify — for humans and agents alike. |
| na | Machine-readable, versioned contract | —/25 | No surface produced evidence for this capability in this run. | A current OpenAPI version with a declared versioning scheme lets agents reason about the contract. |
| na | Auth declared & discoverable | —/25 | No surface produced evidence for this capability in this run. | Agents can only authenticate when auth is declared, scoped, and discoverable. |
Developer ExperienceThe context both developers and agents need to integrate fast — onboarding, code samples, complete descriptions and worked examples2 pass0 warn2 fail59D
| Signal | Points | Findings | Rationale | |
|---|---|---|---|---|
| pass | Quickstart presentvia docs | 25/25 | Quickstart at https://developer.shell.com/docs/welcome-shell-developer-portal includes a runnable first-call code sample. Investigated: docs 100%. | A quickstart is the fastest path from landing page to first successful call. |
| pass | Description completenessvia sdk | 15/15 | 5/5 operations carry a description. Investigated: sdk 100%. | Complete descriptions are the context humans and agents need to use endpoints. |
| fail | Self-service developer portalvia docs | 0/29 | The signup page is sales-gated (contact-sales / request-access language detected), so an agent can't self-serve. Investigated: docs 0%. | Self-serve key/account creation is the fast first call for partners, with no sales gate. |
| fail | Code samples in docsvia docs | 0/18 | No code samples detected across 3 sampled docs pages, so a coding agent gets no ready-to-use examples. Investigated: docs 0%. | Multi-language samples shorten time-to-first-call. |
| na | Changelog published | —/13 | No surface produced evidence for this capability in this run. | A published changelog lets partners track changes without surprise. |
Agent DiscoveryPartners and their agents can find your APIs — llms.txt, registries, crawlable and reachable docs2 pass1 warn1 fail70B
| Signal | Points | Findings | Rationale | |
|---|---|---|---|---|
| pass | Registry & SDK presencevia docs | 28/28 | Indexed on Context7 (moudey/shell, 723 snippets). Investigated: docs 100%, sdk 67%, cli 0%. | Listing in MCP registries and publishing SDKs puts the API where agents and their tooling look. |
| warn | Docs reachable, not hard auth-gatedvia docs | 22.5/30 | Server ignores Accept: text/markdown header (0/1 pages return markdown). Investigated: docs 75%. | Agents can only index and fetch docs they can reach — past auth gates and over correct HTTP semantics. |
| pass | Crawlable / AEOvia wellknown | 12/12 | robots.txt lets all monitored AI crawlers reach the docs paths. Investigated: wellknown 100%. | Bots allowed plus a fresh sitemap make docs findable by agent crawlers. |
| fail | llms.txt present, valid & comprehensivevia docs | 0/30 | No llms.txt found at any candidate location (https://developer.shell.com/llms.txt, https://developer.shell.com/docs/llms.txt). Investigated: docs 0%, sdk 0%. | A valid, comprehensive llms.txt is the machine-readable entry point for agents. |
Agent UnderstandingAgents can correctly interpret your APIs — machine-readable errors, consistent descriptions, structured data, parseable docs1 pass3 warn1 fail73B
| Signal | Points | Findings | Rationale | |
|---|---|---|---|---|
| pass | Machine-readable errors (RFC 9457)via sdk | 28/28 | Single base error "ShellException" with 3 subclasses. Investigated: sdk 100%. | RFC 9457 problem details and a documented error-code inventory let agents parse failures without burning tokens. |
| warn | Agent-navigable, token-efficient docsvia docs | 17.3/22 | No pages support .md URLs (0/1 tested). Investigated: docs 79%. | Server-rendered, clean, small-footprint docs are what an agent can cheaply fetch and parse correctly. |
| warn | Operation purpose clarityvia sdk | 12.5/25 | Read/write classification is mostly ambiguous: only 0/5 operations classified. Investigated: sdk 50%. | Agents select the right endpoint from its summary + operationId; clear, named operations make tool-selection reliable — the strongest driver of correct tool choice. |
| warn | Agent instructions file (AGENTS.md)via sdk | 5/10 | No agents.md or skill.md context file found (nice-to-have for agent operation). Investigated: sdk 50%, wellknown 0%. | An AGENTS.md gives coding agents explicit setup, auth, and usage instructions to interpret and operate the API — beyond llms.txt's link index. |
| fail | Docs structured datavia docs | 0/8 | No JSON-LD or OpenGraph/meta tags found across 2 assessed pages (2 JS-rendered), so answer engines have nothing to cite. Investigated: docs 0%. | Structured data (JSON-LD/schema.org) on docs pages gives agents an unambiguous parse target and is what answer engines cite. Detected on the JS-rendered head (Firecrawl) for a bounded page budget, so JS-injected JSON-LD is now caught; pages we can't render are excluded rather than failed. |
| na | Description consistency across surfaces | —/7 | Only 1 surface description(s) with ≥6 tokens available; need at least 2 to compare. | Every surface tells the same story about what the product is. |
Agent UsabilityAgents have the context to use your APIs reliably, not just find them2 pass1 warn0 fail85A
| Signal | Points | Findings | Rationale | |
|---|---|---|---|---|
| pass | Idempotency documentedvia sdk | 27/27 | Built-in retry machinery is present (grep-derived). Investigated: sdk 100%. | Documented idempotency lets agents retry safely. |
| warn | Sandbox separationvia sdk | 10/20 | No test-mode flag or sandbox environment is exposed. Investigated: sdk 50%. | An isolated environment lets agents exercise destructive operations safely. |
| pass | Runnable collection with test scriptsvia sdk | 9/9 | README is present and includes a runnable quickstart. Investigated: sdk 100%, platform 0%. | A public, maintained collection with assertions is runnable truth agents validate against. |
| na | Rate-limit signaling | —/22 | No surface produced evidence for this capability in this run. | Machine-readable rate-limit headers let agents throttle adaptively. |
| na | Pagination documented & consistent | —/22 | No surface produced evidence for this capability in this run. | Consistent, documented pagination lets agents traverse collections. |
Resources Discovered
The public resources we found for Shell — the evidence behind the score. All discovered from public sources; nothing here requires access to your systems.
| Agent hints | Context7 (723) |
|---|