Colter Check

TL;DR: Run colter check https://your-store.com for a free, read-only readiness scan. Add --json when an agent, script, or CI job needs structured output.

colter.check and colter check --json return structured JSON optimized for AI agent and automation workflows.

What Check Measures

Check scores five dimensions:

Dimension	Weight	Focus
Transaction	30%	UCP, ACP, checkout readiness, PSP compatibility
Discovery	25%	JSON-LD, sitemap, llms.txt, ai-plugin.json, robots.txt, Open Graph
Ecosystem	20%	Google and OpenAI ecosystem coverage
Content Quality	15%	Product data completeness and structure
Security	10%	HTTPS, headers, rate limiting, auth signals

See Scoring Model for the full signal map.

Verdicts

Verdict	Meaning
AGENT-READY	Public UCP and ACP evidence detected. Runtime catalog, cart, checkout, and payment behavior still require Test before purchase-readiness claims.
PARTIALLY AGENT-READY	One protocol or major public readiness path detected. Missing protocols or runtime paths remain.
NOT AGENT-READY	No public commerce protocol evidence detected. Ordinary web crawling may still work, but structured AI commerce coverage is not verified.

Evidence Taxonomy

Check separates public evidence from assumptions:

Status	Meaning
verified	Colter directly observed the signal on public pages or endpoints.
inferred	Colter detected a platform or storefront pattern that suggests coverage, but did not verify the underlying runtime behavior.
runtime_test_required	The requirement depends on live agent, cart, checkout, payment, or authenticated behavior and belongs in Test.
gap	Colter looked for the signal and did not find it.
not_checked	The quick scan did not have enough public evidence to evaluate the field.

CLI

colter check <url> [flags]

Common Flags

Flag	Purpose
--json	Machine-readable output
-v, --verbose	Detailed probe output
--batch FILE	Batch mode with YAML input
--concurrency N	Parallel batch workers
--ci	Exit non-zero below the threshold
--threshold N	Score floor for --ci
--api-url URL	Override the API base URL

Examples

colter check example-store.com

colter check example-store.com --json

colter check example-store.com -v

colter check example-store.com --ci --threshold 70

Batch Mode

colter check --batch stores.yaml --concurrency 10 --json

stores:
  - url: store-a.com
  - url: store-b.com

For text and CSV-oriented portfolio runs, colter batch is often the simpler fit.

Output Highlights

Key fields in --json output:

Field	Meaning
protocols	UCP, ACP, and MCP detection state
coverage	Google and OpenAI ecosystem reach
scores	Per-dimension details and fixes
composite_score	Final score from 0 to 100
agentic_level	Maturity level derived from core gates
verdict	Canonical readiness state
evidence_summary	Counts and notes separating verified, inferred, runtime-test, and gap evidence
catalog_readiness	Product catalog evidence for names, price/offers, availability, images, identifiers, and agent catalog surfaces
agent_handoff	Next-action URLs and concise handoff text for Fix and Lens

What Gets Probed

• discovery files and metadata
• protocol endpoints
• product and payment signals
• security headers
• latency and responsiveness

Typical Next Move

• low discovery score: run Fix
• protocol issues: run Verify
• interaction gaps: run Test
• live traffic questions: set up Lens

Next Steps

• Scoring Model
• Fix
• Lens
• Evidence Packs
• API Reference