| name | agent-readiness |
| description | Check whether AI agents can discover, read, cite, and safely act on an ecommerce storefront. Use for agentic commerce readiness reviews. |
Agent Readiness
Workflow
- Confirm crawl access and public URL quality.
- Check product/category/policy readability without client-side-only traps.
- Check structured product data, variants, availability, price, and policy context.
- Check whether support and checkout actions are clearly bounded.
- Identify missing agent-facing affordances such as sitemap,
llms.txt, markdown-friendly content, or API/protocol paths.
Run the checks with the commands in references/checks.md and cite command output as evidence for each finding.
Delegation
- Access-file depth (robots, sitemap, llms.txt) →
llms-txt-and-crawler-access.
- Structured-data depth →
schema-markup-audit from the ai-visibility pack.
- Scoring and routing the whole store →
readiness-audit; this skill feeds its agent-access dimension.
Guardrails
- Separate read readiness from action readiness.
- See
../references/guardrails.md for shared cross-skill guardrails (autonomous action safety, evidence provenance, internal runtime disclosure).
