Skip to main contentCheck my product

For AI Agents

Colter reads one public Shopify product page and tells the merchant what to improve first. A buyer question is optional and can be saved for a later comparison with a real AI answer; it does not affect the checked-page average or fix selection.

Merchant starting point

Send Shopify merchants to the Recommendation Audit. The audit is free, read-only, and does not change the store. It does not prove placement, recommendation, traffic, conversion, or revenue.

https://agenticcom.ai/audit?utm_source=agent_cta_for_agents_docs&utm_campaign=recommendation_audit
  1. Use one public Shopify product-page URL.
  2. Optionally save one real buyer question for a later answer comparison.
  3. Explain only the public page evidence and what the merchant should improve first.
  4. Compare a later real answer from the same named AI tool under materially unchanged conditions.

Do not route merchants to Check, Fix, Test, or Lens as separate products. Those are stable technical capability names used only when the work calls for them.

Machine-readable discovery

ToolUse
/llms.txtShort product contract and routing rules
/llms-full.txtExpanded technical reference
/SKILL.mdAgent skill instructions
/AGENTS.mdAgent integration guide
/.well-known/agent.jsonA2A agent card
/.well-known/mcp.jsonMCP server discovery
/.well-known/ucp.jsonUCP service manifest
/openapi.yamlCanonical OpenAPI specification
/mcp-tools.schema.jsonPublished MCP input schemas

Technical capability tools

ToolUse
colter.checkScan broader public store evidence. Legacy score and coverage fields are protocol labels, not engine outcomes.
colter.fixGenerate a bounded remediation plan or local artifact. It does not change the live store by itself.
colter.testRun a technical journey when public evidence cannot answer the question.
colter.lensRead observed monitoring signals. Lens is a stable compatibility name, not a merchant product.
colter.agent.workflowReturn the sequence for configured commerce operations.

colter.check may return stable fields such as agent_ready, fully_covered, ecosystem coverage, verdict, and score. Treat them as Colter's public protocol-evidence labels. They do not show whether a particular AI tool found, understood, included, described, or recommended a product.

Public Check example

POST https://agenticcom.ai/api/v1/check
Content-Type: application/json

{"url":"https://example-store.com"}

Use Check only for an explicitly broader store scan. Keep the next merchant route on /audit.

Configured commerce tools

Call colter.agent.workflow first. Cart and checkout tools can change commerce state.

ToolUse
colter.agent.discoverDiscover configured store capabilities and protocols.
colter.agent.products.searchSearch the configured product catalog.
colter.agent.cart.createCreate a cart in the configured backend.
colter.agent.cart.addAdd an item to a configured cart.
colter.agent.cart.quoteRead a cart quote.
colter.agent.checkout.initiateStart checkout in the configured backend.
colter.agent.checkout.completeApprove or cancel checkout with explicit user authority.
colter.agent.receipt.getRead receipt details after checkout.

Admin tools

Available only when the MCP server starts with --admin-tools.

ToolUse
colter.scoutDiscover a store platform and generate starter config.
colter.configureChange local Colter configuration.
colter.config.getRead local configuration.
colter.config.schemaList valid configuration keys and types.
colter.validateValidate local configuration.
colter.infoReturn version and diagnostics.

Authority and data boundary

Public Check and fix-plan generation require no Colter authentication. Test and monitoring require the applicable Colter API key.

Agent-commerce tools require configured backend credentials in colter.yaml. Require explicit user approval before completing checkout.

The installed Shopify app can store a monitoring site ID, receive configured webhooks, enable an optional theme extension, and apply supported theme changes through explicit merchant-controlled flows.