# AgentCheck

> AgentCheck (agentcheck.care) is an AI-agent diagnostic service: give it the URL of a bot — an A2A agent or an OpenAI-compatible chat endpoint — and it runs synthetic-persona conversations, prompt-injection and PII-leakage tests, hallucination and brand-alignment checks, and returns a scored report by magic link. Free scan with no login; paid one-off tiers at $10, $25 and $75 via Stripe.

This file was generated by API Evangelist from AgentCheck's public site and served documents on 2026-09-19. AgentCheck publishes no llms.txt of its own (https://agentcheck.care/llms.txt returns 404). Every link below was fetched and returned the status noted.

## Core Pages

- [Home](https://agentcheck.care/): Product overview, tier pricing (#pricing), the checkup form, and the ten diagnostic modules. 200.
- [Terms of Service](https://agentcheck.care/terms): Free-scan limits (30/week pool, 2/week per user), refund policy, 5-minute connection window, Delaware governing law. 200.
- [Privacy Policy](https://agentcheck.care/privacy): What is collected, Stripe/Cloudflare/Google Gemini as processors, report retention and deletion within 7 business days, GDPR rights. 200.

## Machine-Readable Surface

- [OpenAPI 3.1.0](https://agentcheck.care/openapi.json): 35 operations, FastAPI-generated, no securitySchemes declared. 200, application/json.
- [Swagger UI](https://agentcheck.care/docs): Interactive reference rendered from the OpenAPI. 200.
- [ReDoc](https://agentcheck.care/redoc): Read-only reference rendered from the OpenAPI. 200.
- [A2A Agent Card](https://agentcheck.care/.well-known/agent-card.json): A2A 0.3.0 card, two skills (free-scan, paid-checkup), endpoint https://agentcheck.care/a2a, anonymous JSON-RPC. 200.
- [Tier catalog](https://agentcheck.care/api/tiers): JSON of the four tiers with price and included quotas. 200.
- [Free-scan pool](https://agentcheck.care/api/free-scans): Live weekly pool counter - used, max, remaining, reset_in_seconds. 200.
- [Health](https://agentcheck.care/api/health): {"status":"ok","version":"0.1.0"}. 200.
- [robots.txt](https://agentcheck.care/robots.txt): Disallows /api/, /checkup/, /report/ for crawlers. 200.

## How an agent uses it

- A2A: POST JSON-RPC 2.0 message/send to https://agentcheck.care/a2a with text such as "Run a free scan on https://my-bot.example.com" or "Test https://my-bot.example.com with Full Check" (the paid skill returns a payment link).
- REST: GET /api/free-scans to check quota, POST /api/validate-bot to confirm the target is reachable, POST /api/checkup (tier "free") or POST /api/checkout (tier basic|full|enterprise), then GET /api/checkup/{checkup_id}/stream (SSE) and GET /api/checkup/{checkup_id}/report.
- Exam mode: instead of exposing a bot URL, the customer points their bot at a generated OpenAI-compatible endpoint POST /exam/{token}/v1/chat/completions and AgentCheck drives the conversation from the other side.

## Not published

- No SDKs, CLI, MCP server, status page, changelog, security.txt, OAuth metadata, APIs.json or api-catalog (all probed 2026-09-19; see the API Evangelist profile at https://github.com/api-evangelist/agentcheck-care).
- Contact: hello (at) agentcheck.care (site footer).