Amazon
amazon.com · E-commerce / D2C Retail modeTransactional
Last probed 2026-08-09 23:51 UTC · Provisional
Amazon — General e-commerce marketplace.
Verdict: Partially — an AI agent can use Amazon, but activate is blocked — A CAPTCHA blocks automated signup.
- DiscoverNot assessed
- ActivateBlocked
- OperateNot assessedgated by Activate
- ManageNot assessedgated by Activate
Checkout flow not yet probed (Tier 2). For a transactional listing the purchase path is where interrupts actually bite — this readout reflects only what passive and signup-level checks could observe, not a verified purchase.
Capability parity
Each stage answers one question about what an agent can do unattended. Stages are ordered by the credential-gateway dependency: Activate issues the keys that unlock Operate and Manage.
Discover
Can an agent find and read it, no account?
- Be found by a retrieval agentNot assessed
- Read what it does and how to use itNot assessed
- Read pricing & limits without loginNot assessed
Activate
Can an agent sign up and walk away with the keys it needs?
- Create an account self-serveBlocked
A CAPTCHA blocks automated signup.
CAPTCHA / bot-challenge blocks this.
Evidence
- Path tried
- signup flow
- Where it stalled
- verification step
- Evidence tier
- Tier 0 (passive scan)
- Verification
- Evidence-only — graded from passive signal, not live-verified.
- CAPTCHA / bot-challengeBefore any account existsPlain-HTTP fetch bot-walled (HTTP 503, bot-mitigation page) while a real browser loads fine - non-browser agents are blocked outright.
- Obtain an API / MCP access credentialEasy
Self-serve credential path detected.
Evidence
- Path tried
- docs → key issuance
- Where it stalled
- completed
- Evidence tier
- Tier 1 (unauth probe)
- Verification
- Evidence-only — graded from passive signal, not live-verified.
- Backing checks
api_key_self_serve
- Obtain a management / admin keyNot assessed
- Accept the terms without a bot-banEasy
No bot-prohibition signal found.
Evidence
- Path tried
- ToS / policy scan
- Where it stalled
- completed
- Evidence tier
- Tier 0 (passive scan)
- Verification
- Evidence-only — graded from passive signal, not live-verified.
Operate
Can an agent do the service's actual jobs — call, transact, pay?
Gated by Activate — the credential this stage needs isn't issued unattended, so these rows may be unreachable even where they read Easy.
- Use the service's primary jobNot assessed
Gated by Activate — the credential this needs isn't issued unattended.
- Complete a transaction / checkoutNot assessed
Live checkout not yet attempted (Tier 2).
Gated by Activate — the credential this needs isn't issued unattended.
- Pay via an agent rail (x402 / ACP / AP2)Not assessed
Gated by Activate — the credential this needs isn't issued unattended.
Manage
Can an agent administer the account and provision resources?
Gated by Activate — the credential this stage needs isn't issued unattended, so these rows may be unreachable even where they read Easy.
- Provision / scale / tear down a serviceNot assessed
Gated by Activate — the credential this needs isn't issued unattended.
- Change account details / set budget / rotate keysNot assessed
Gated by Activate — the credential this needs isn't issued unattended.
- View usage & billingNot assessed
Gated by Activate — the credential this needs isn't issued unattended.
- Cancel / export / delete (offboard)Not assessed
Gated by Activate — the credential this needs isn't issued unattended.
Findings
How scoring works →Interrupt tier: High-Interrupt — derived from 2 findings; interrupt score 44 / 100 — the badge-driving metric. (range 13–63 · 9 unassessed)
Every tool starts at 100. Each interrupt found subtracts a penalty of base weight × timing × hardness × evidence — the multiplier values below come from the same rubric tables the score is computed with. New snapshots also provisionally subtract 50% of the full worst-case tax for each unassessed property and 25% for evidence-graded successes not live-verified by Tier 2.
Property coverage
A check only counts as clean when a probe tier capable of detecting that interrupt actually ran and found nothing. Not tested ≠ clean.
| Property | Stage | State | Evidence | Required |
|---|---|---|---|---|
| CAPTCHA / bot-challenge | activate | failed | Active finding — itemized in the penalty matrix below. Tier 0 assesses this property. | no |
| SMS / phone OTP verification | activate | failed | Active finding — itemized in the penalty matrix below. Tier 1 assesses this property. | no |
| Email OTP verification | activate | unassessed | Tier 1 (unauth probe) ran but did not gather enough evidence to claim clean. Tier 1 assesses this property. | no |
| Mandatory phone verification call | activate | unassessed | No probe tier that can detect this has run yet — unknown, not clean. Tier 2 assesses this property. | no |
| Mailed physical code (postal) | activate | unassessed | No probe tier that can detect this has run yet — unknown, not clean. Tier 2 assesses this property. | no |
| “Contact sales” wall (no self-serve path) | activate | unassessed | No probe tier that can detect this has run yet — unknown, not clean. Tier 2 assesses this property. | no |
| Manual approval queue | activate | unassessed | No probe tier that can detect this has run yet — unknown, not clean. Tier 2 assesses this property. | no |
| Rate-limit-triggered manual review | manage | unassessed | No probe tier that can detect this has run yet — unknown, not clean. Tier 2 assesses this property. | no |
| One-time account / payment-method provisioning | manage | unassessed | No probe tier that can detect this has run yet — unknown, not clean. Tier 2 assesses this property. | no |
| Identity / KYC verification | activate | unassessed | No probe tier that can detect this has run yet — unknown, not clean. Tier 2 assesses this property. | no |
| Crypto wallet / key-custody setup | operate | unassessed | No probe tier that can detect this has run yet — unknown, not clean. Tier 2 assesses this property. | no |
| ToS clause banning automated/bot access | discover | success | None found — checked at Tier 0 (passive scan) (evidence-graded; not yet live-verified by a Tier 2 probe) Tier 0 assesses this property. | no |
| Mandatory redirect to vendor’s own site | operate | success | None found — checked at Tier 1 (unauth probe) (evidence-graded; not yet live-verified by a Tier 2 probe) Tier 1 assesses this property. | no |
| SSO-only signup, no API-key issuance | activate | success | None found — checked at Tier 1 (unauth probe) (evidence-graded; not yet live-verified by a Tier 2 probe) Tier 1 assesses this property. | no |
| No Machine Interface | operate | success | None found — checked at Tier 1 (unauth probe) (evidence-graded; not yet live-verified by a Tier 2 probe) Tier 1 assesses this property. | yes |
| Interrupt found | Base weight | × Timing | × Hardness | × Evidence | = Penalty |
|---|---|---|---|---|---|
| Starting score | 100 | ||||
| CAPTCHA / bot-challengeBefore any account existsPlain-HTTP fetch bot-walled (HTTP 503, bot-mitigation page) while a real browser loads fine - non-browser agents are blocked outright. | 6 | ×3ongoing | ×2.5hard | ×0.6Tier 0 (passive scan) | −27 |
| SMS / phone OTP verificationAccount verification (KYC / OTP)Signup form collects a phone/SMS number — indicates SMS/phone verification during onboarding. | 5 | ×1one-time setup | ×2.5hard | ×0.8Tier 1 (unauth probe) | −10 |
| Total penalty | −37 | ||||
| Full unknown tax 98.2 — capped at 25 (point score uses 50% of the capped value) | −98.2 | ||||
| Evidence-graded tax (successes not live-verified by a Tier 2 probe) 139.2 — capped at 25 (point score uses 25% of the capped value) | −139.2 | ||||
| Raw score = 100 − 37 = 63 Point score = 63 − 50% × 25 − 25% × 25 = 44.3 → 44 | 44 | ||||
What would make this Easy
One concrete change per blocker found. These are the diffs between this listing's current parity and an unattended agent path.
- Activate — CAPTCHA / bot-challengeDrop the CAPTCHA, or gate it behind a documented bot-detection signal a well-behaved agent can pass.
- Activate — SMS / phone OTP verificationAllow email-only verification, or offer SSO alongside an API-key path.
How we scored this
- Confidence: Provisional (0.55) — Based on Tier 0/1 evidence only (passive checks, no live transaction attempt) — treat the number as a triage signal, not a verified outcome.
- Machine-Readability sub-score: 29 / 100 (a separate integration-ease signal — it never enters the Interrupt Score or badge).
- Some capability stages are not yet assessed — shown plainly in the spine above, never as a pass.
- Rubric, tiers, confidence model, and badge thresholds →
Machine-readability checks 29 / 100
Four check groups, equal weight: machine interfaces, browsing efficiency, element discoverability, and discovery/agent access. Each check carries an evidence grade (proven / plausible / emerging) — an honesty label that never weights the score. What these mean →
- llms.txtemergingNot provided
- llms-full.txtemergingNot provided
- AGENTS.mdemergingNot provided
- A2A agent cardemergingNot provided
- MCP server (well-known or official registry)plausibleNot provided
- OpenAPI/Swagger specplausibleNot provided
- Rate-limit headersplausibleNot provided
- Structured JSON error responsesplausibleNot provided
- OAuth authorization-server metadata (RFC 8414)provenNot provided
- API catalog (RFC 9727) (emerging)emergingNot provided
- Agent-payment discovery (x402 / UCP) (emerging)emergingNot provided
- Lean HTML transfer sizeplausible4 KB raw HTML (threshold 488 KB)
- Manageable DOM sizeplausible39 DOM nodes (threshold 1500)
- High content-to-markup densityplausible3.8% text/HTML (142 chars in 4 KB, threshold 2%)
- Content readable without JS executionprovenraw HTML holds 97% of rendered text (142 vs 146 chars; threshold 40%)
- Clean-content alternate (RSS/Atom/feed/markdown)plausibleNot provided
- Serves markdown on Accept: text/markdownplausibleNot provided
- Interactive elements are semanticplausible3/3 semantic (100%); 0 div/span click targets
- Form fields have accessible namesplausibleNot checked — no form fields on the representative page
- Inputs carry autocomplete hintsplausibleNot checked — no autocomplete-eligible inputs
- WebMCP tools (modelContext API)emergingNot provided
- AI retrieval/search fetchers allowed (robots.txt)provenNot provided
- Sitemap (indexability)plausibleNot checked — sitemap lookup unreachable
- schema.org JSON-LD in raw HTMLplausibleNot provided
- Content-Signal directives (robots.txt)emergingNot provided
- Web Bot Auth key directory (emerging)emergingNot provided
- Discovery Link headers on homepage (emerging)emergingNot provided
- robots.txt AI-crawler accessplausibleNot provided
- Unauthenticated API error is structured JSONplausibleNot provided
- Unauthenticated API returns auth guidance (401/403 or documented)plausibleFound — api.amazon.com/ (HTTP 403)
- API exposes rate-limit headersplausibleNot provided
- API key obtainable without human interactionplausibleAPI documentation host reachable: https://docs.amazon.com/, https://developer.amazon.com/
Evidence & history
Last probed 2026-08-09 23:51 UTC. The raw JSON is the archival record.
| Date | Score | Badge | Rubric version |
|---|---|---|---|
| 2026-08-09 23:51 UTC | 44 | High-Interrupt (current) | [email protected] |
| 2026-08-09 23:51 UTC | 54 | Moderate-Interrupt | [email protected] |
| 2026-08-09 17:13 UTC | 81 | Low-Interrupt | [email protected] |
| 2026-08-09 05:28 UTC | 88 | Low-Interrupt | [email protected] |
| 2026-08-09 05:28 UTC | 88 | Low-Interrupt | [email protected] |
| 2026-08-09 03:14 UTC | 88 | Low-Interrupt | [email protected] |
| 2026-08-09 01:50 UTC | 88 | Low-Interrupt | [email protected] |
| 2026-08-09 01:45 UTC | 88 | Low-Interrupt | [email protected] |
| 2026-07-10 07:19 UTC | 100 | Low-Interrupt | [email protected] |
Alternatives in E-commerce / D2C Retail
Other E-commerce / D2C Retail tools in the same mode (transactional), ranked by Overall LLM-Friendliness (the mean of Interrupt Score and Machine-Readability) — comparable to this listing because they share its usage mode.