# AgentReady.me — agent journey audits and verified fixes

## Purpose

AgentReady evaluates whether public websites and connected repositories are discoverable, understandable, operable, and safe for AI agents, then returns evidence-backed fixes and verification steps.

For broader launch-readiness checks beyond agent visibility, use [LaunchReady](https://launchready.me).

## When to use AgentReady

- Before launch or after a major site change, scan a public URL for agent-readiness gaps.
- When you need a prioritized Fix Pack with evidence, candidate targets, review gates, and fix-specific verification.
- After fixes are deployed, call `verify_fixes` to check stable finding IDs rather than relying on score movement alone.

Do not use AgentReady as a security certification, legal opinion, accessibility conformance report, or guarantee of search ranking. Do not scan private or local-network URLs.

## Safe public actions

- Read the methodology and resource library.
- Run a public URL scan at `POST /api/v1/scan` with JSON such as `{ "url": "example.com" }`. Scans are not published by default; set `publishReport: true` only after the user explicitly asks for a shareable link.
- Retrieve an opted-in report at `GET /api/v1/report?id=<opaque-report-id>`; add `history=true` for its bounded history.
- Check whether paid checkout is available at `GET /api/checkout`.
- Download the AgentReady audit skill from `/skills/agentready-audit/SKILL.md`.
- Call the public MCP endpoint at `/api/mcp` with no authentication; it reads public web content and exposes no private account data or delegated write actions.
- The MCP tools include `get_public_report` for an opted-in report id or its history.

## Actions that require the user

- Account creation and sign-in.
- Connecting a GitHub or Vercel account.
- Starting Stripe Checkout or changing a subscription.
- Sharing a private report or repository.

Never enter payment details, grant third-party account access, or complete an irreversible action without explicit user confirmation.

## Canonical resources

- Product: https://www.agentready.me/
- Methodology: https://www.agentready.me/methodology
- Resources: https://www.agentready.me/resources
- llms.txt checker and generator: https://www.agentready.me/tools/llms-txt
- AI visibility vs. agent readiness: https://www.agentready.me/ai-visibility-vs-agent-readiness
- Pricing: https://www.agentready.me/pricing
- OpenAPI: https://www.agentready.me/openapi.json
- API catalog (RFC 9727): https://www.agentready.me/.well-known/api-catalog
- API deprecation policy: https://www.agentready.me/deprecation-policy.md
- Developer documentation: https://www.agentready.me/developers
- Agent manifest: https://www.agentready.me/.well-known/ai-agent.json

The public scan and MCP surfaces intentionally require no authentication because they only read public content. Shareable reports require explicit opt-in, use opaque unlisted ids, expire after 30 days, and contain public evidence only. Future protected resources may require narrowly scoped authorization.