Understanding Your Colter Report
A Colter Check scores your store across 30+ signals and 5 weighted dimensions. The average store scores around 30/100, so the report exists to show where agents can work today and where they stop.
Foundation Signals
These checks tell you whether agents can discover, read, and trust your storefront before any advanced protocol comes into play.
JSON-LD and product fields
Colter checks whether Product schema exists, whether it is valid, and whether important fields like price, availability, brand, images, and offers are complete enough for agents to act on.
Sitemap and robots posture
Colter verifies that your sitemap loads, exposes products, and that robots.txt allows storefront crawling while still protecting admin and checkout paths.
Open Graph and SSR readiness
The report checks whether product pages have usable OG tags and whether product content is available in server-rendered HTML rather than hidden behind client-only rendering.
Agent guidance files
Colter looks for llms.txt, AGENTS.md, and Markdown-for-Agents because they give models explicit instructions, navigation hints, and machine-friendly context.
Deep discovery probes
The report also checks AI crawler posture, ai-plugin.json, UCP manifest validity, and whether public endpoints respond cleanly enough for agents to continue.
Security and traffic controls
Security checks cover HSTS, CSP, X-Frame-Options, X-Content-Type-Options, rate-limit headers, and agent auth support where relevant.
AI Agent Compatibility
Protocol results show which ecosystems can do more than crawl. They indicate where agents can discover tools, move through a transaction, or hand work off to another agent.
UCP -- Universal Commerce Protocol
Checks for a manifest at /.well-known/ucp, validates the response, and inspects whether the store exposes browse, cart, checkout, and payment capabilities.
ACP -- Agent Commerce Protocol
Checks whether the store exposes an ACP-compatible endpoint for conversational product discovery, order handling, and Stripe-oriented payment flows.
MCP -- Model Context Protocol
Checks for MCP server availability or structured tool affordances that let Claude-style agents interact with store services intentionally.
WebMCP -- browser-native agent tools
Checks for WebMCP HTML markers, tool counts, checkout tool availability, parameter descriptions, and autosubmit support on important forms.
A2A -- agent-to-agent discovery
Checks for an agent card at /.well-known/agent.json, validates the card, and inspects declared capabilities and skills.
Scoring Dimensions
The composite score is not a flat checklist. It is a weighted model designed around how agents actually shop.
Discovery -- 25%
JSON-LD, sitemap, robots.txt, OG tags, llms.txt, ai-plugin.json, AGENTS.md, Markdown-for-Agents, UCP manifest validity, and AI crawler posture.
Transaction -- 30%
UCP browse/cart/checkout/payment capabilities, ACP support, Shopify MCP, endpoint responsiveness, and payment-service-provider detection.
Security -- 10%
HSTS, CSP, X-Frame-Options, X-Content-Type-Options, rate-limit headers, and agent auth support.
Ecosystem -- 20%
Coverage across UCP, ACP, MCP, WebMCP, A2A, plus ecosystem indicators like x402 support and agent guidance.
Content Quality -- 15%
JSON-LD field completeness, offer quality, OG quality, pricing accuracy, availability accuracy, and SSR readiness.
Verdict Levels
After those checks, Colter assigns a verdict. That verdict tells you how much confidence an agent should have right now.
Agent-Ready
The store clears the core foundations and exposes enough protocol or browser affordance for agents to act with confidence. This is where Colter Lens becomes useful for regression tracking after launch.
Partially Agent-Ready
Agents can find parts of the store, but key gaps remain: thin product data, missing protocol coverage, weak security posture, or brittle browser flows. This is where most stores land.
Not Agent-Ready
Critical discovery or content issues prevent agents from understanding or trusting the store. Start with the highest impact foundation fixes before thinking about protocol depth.
Plain Language Summary
Every report includes a plain-language summary so a merchant, developer, or agency partner can understand the outcome without parsing every signal.
If you need behavioral proof, Colter Test runs 11 AI personas through real shopping journeys on your store. That includes personas like Gemini Shopper, Operator Shopper, and Security Shopper, which helps explain why a technical score may still feel weak in practice.
What to do based on your verdict
If you are Agent-Ready:
Keep monitoring. Colter Lens will catch score drops after theme changes, app installs, protocol regressions, or catalog updates.
If you are Partially Agent-Ready:
Fix the highest-impact gaps first. Colter Fix is useful when you need platform-specific code for missing schema, manifests, or browser annotations.
If you are Not Agent-Ready:
Start with the platform guides for Shopify or WooCommerce. For custom stacks, use the discoverability guide and rebuild from the foundation layer up.
You're already listed. Get the report.
AI agents are already evaluating your store. Run a Colter Check and see your score across 30+ signals and 5 dimensions.
See What Agents See