Benchmark your endpoints for AI Agent execution
Audit your domain against the 22 core Agent-Readiness Protocol standards (Level 0 through Level 5). Discover where autonomous runtimes stall, where tokens leak, and how to reach Level 4/5 integration.
Telemetry generated against global edge protocol benchmarks (Level 0 through Level 5). Architectural remediation engineered by TENSIX.
Edge Diagnostic Architecture
The TENSIX Agent-Readiness Diagnostic (ARD-22)™ benchmarks your digital surface against 22 ratified machine-interaction specifications—spanning RFC 9421 cryptographic request signing, RFC 9727 API discovery catalogs, WebMCP tool declarations, and Content-Signals content negotiation. Telemetry executes client-side in real time with zero external data retention.
While generic security firewalls flag bots indiscriminately, TENSIX delivers architectural protocol triage: distinguishing malicious scrapers from legitimate autonomous agent execution (Claude in Chrome, OpenAI Operator, Meta Muse).
The Agent-Readiness Protocol Spectrum
AI assistants and autonomous agents execute tasks by reading structured web endpoints. If your protocol rails fail, they bypass your service and transact with competitors.
The diagnostic benchmarks endpoints across 6 defined capability tiers:
- Level 0 — Not Ready: Missing essential crawler discovery headers and canonical signals.
- Level 1 — Basic Web Presence: Standard robots.txt, XML sitemap, and RFC 8288 link relations.
- Level 2 — Bot-Aware: Granular AI bot directives and Content-Signals content declaration.
- Level 3 — Agent-Readable: Automated content negotiation serving deterministic markdown/plain-text variants.
- Level 4 — Agent-Integrated: Machine-readable RFC 9727 API catalogs, WebMCP tool manifests, and agent skill cards.
- Level 5 — Agent-Native: Cryptographically signed requests (RFC 9421), authenticated machine identities, and zero-OTP transactional rails.
Commercial Moat: Level 4 & Level 5 Compliance
Over 90% of business websites stall at Level 2. Reaching Level 4 (Agent-Integrated) and Level 5 (Agent-Native) unlocks machine-to-machine commerce: exposing sovereign WebMCP tool manifests, authenticating AI transactions via RFC 9421 signatures, and automating checkout flows.
Schedule Enterprise Protocol Remediation →Protocol Architecture FAQ
What does the TENSIX Agent-Readiness Diagnostic test?
The diagnostic evaluates 22 technical standards across five core layers: Autonomous Discoverability, Machine-Readable Content Negotiation, AI Bot Access Control & Request Signing, API/Skill Exposure (WebMCP & RFC 9727), and Machine-to-Machine Commerce Readiness.
Does TENSIX store my website address or my results?
No. The diagnostic executes client-side with zero data retention. Your endpoint URLs and telemetry never persist on external databases.
Why do some checks fail even on high-traffic websites?
Over 90% of business websites stall at Level 2 (Bot-Aware). Advanced checks like RFC 9421 cryptographic signatures, OAuth discovery, and WebMCP agent toolcards require dedicated machine-interaction rails.
Which checks determine whether AI search engines cite your website?
Six foundational protocols dictate citation grounding: robots.txt allowance, XML sitemaps, AI bot crawler access, Content-Signals headers, markdown content negotiation, and published agent skills.
How long does the edge diagnostic take?
Typically 10 to 15 seconds. The engine benchmarks live response latency per protocol check and provides individual duration timings in milliseconds.
Need Protocol Remediation?
Schedule an architecture sprint with TENSIX to implement Level 3 through Level 5 agent rails, RFC 9421 signatures, and WebMCP endpoints.