agentrust-io · imran-siddique · Jul 1, 2026 · Jul 1, 2026 · Jul 1, 2026
diff --git a/.github/workflows/docs.yml b/.github/workflows/docs.yml
@@ -55,7 +55,7 @@ jobs:
 
           if [ -d docs ]; then cp -r docs $BUILD/docs; fi
 
-          for fname in README.md CHANGELOG.md CONTRIBUTING.md GOVERNANCE.md ROADMAP.md LIMITATIONS.md CNAME; do
+          for fname in README.md CHANGELOG.md CONTRIBUTING.md GOVERNANCE.md ROADMAP.md LIMITATIONS.md CNAME robots.txt; do
             if [ -f "$fname" ]; then cp "$fname" "$BUILD/$fname"; fi
           done
 

diff --git a/README.md b/README.md
@@ -26,6 +26,10 @@
 
 > **Developer Preview** - launching at Confidential Computing Summit, June 23 2026. May have breaking changes before v1.0.
 
+**cMCP (Confidential MCP Runtime) is an open-source gateway that enforces MCP tool-call policy inside a hardware Trusted Execution Environment (TEE).** Every tool call is intercepted, evaluated against a Cedar policy bundle, and enforced where the process it governs cannot reach it. Each session produces a signed, hardware-attested TRACE Claim that a verifier checks without trusting the operator.
+
+> **TL;DR** - Point your agent at the cMCP Gateway. It evaluates every tool call against a Cedar policy inside a TEE, blocks or redacts what the policy denies, and emits a tamper-evident TRACE Claim as proof. Run `pip install cmcp-runtime` and start in software mode with no hardware required.
+
 Your agent calls Snowflake, Salesforce, a dozen APIs. What stops it from leaking a customer's data on one of those calls? If a regulator asks, could you prove it didn't?
 
 ---
@@ -241,6 +245,73 @@ See [SECURITY.md](SECURITY.md) for vulnerability reporting and response SLAs. Se
 
 ---
 
+## FAQ
+
+### What is cMCP?
+
+cMCP (Confidential MCP Runtime) is an open-source gateway that enforces MCP tool-call policy inside a hardware Trusted Execution Environment. It intercepts each tool call, evaluates it against a Cedar policy bundle, enforces the decision (allow, deny, or redact), and records the call in a hardware-sealed audit chain.
+
+### How is cMCP different from software-only MCP governance?
+
+Software-only governance runs the policy engine in the same OS an operator or a supply-chain CVE can reach, so it cannot prove the policy that ran was the approved one or that the decision was not flipped in memory. cMCP runs the policy engine inside a TEE and measures the Cedar bundle hash into the hardware attestation report before any code runs, so the control plane cannot be reached by the process it governs.
+
+### Do I need special hardware to try it?
+
+No. Set `CMCP_DEV_MODE=1` to use the software-only TEE provider and run the full quickstart without a hardware TEE. Hardware providers (TPM, AMD SEV-SNP, Intel TDX, OPAQUE) are used in production.
+
+### What is a TRACE Claim?
+
+A TRACE Claim (a `GatewayClaim`) is a signed, hardware-attested artifact produced per session. It records which tools ran, which policy decided each call, the Cedar bundle hash, and the audit chain, and it is signed with an Ed25519 key that never leaves the TEE. A verifier checks it with the `cmcp_verify` library without trusting the operator.
+
+### Which TEE providers are supported?
+
+TPM 2.0 / vTPM, AMD SEV-SNP, and Intel TDX, with NVIDIA GPU confidential computing planned for v0.2 and OPAQUE Confidential Runtime available as explicit opt-in. Auto-detection order is SEV-SNP, then TDX, then TPM, then software.
+
+### What license is cMCP under?
+
+MIT.
+
+<script type="application/ld+json">
+{
+  "@context": "https://schema.org",
+  "@type": "FAQPage",
+  "mainEntity": [
+    {
+      "@type": "Question",
+      "name": "What is cMCP?",
+      "acceptedAnswer": { "@type": "Answer", "text": "cMCP (Confidential MCP Runtime) is an open-source gateway that enforces MCP tool-call policy inside a hardware Trusted Execution Environment. It intercepts each tool call, evaluates it against a Cedar policy bundle, enforces the decision (allow, deny, or redact), and records the call in a hardware-sealed audit chain." }
+    },
+    {
+      "@type": "Question",
+      "name": "How is cMCP different from software-only MCP governance?",
+      "acceptedAnswer": { "@type": "Answer", "text": "Software-only governance runs the policy engine in the same OS an operator or a supply-chain CVE can reach, so it cannot prove the policy that ran was the approved one. cMCP runs the policy engine inside a TEE and measures the Cedar bundle hash into the hardware attestation report before any code runs, so the control plane cannot be reached by the process it governs." }
+    },
+    {
+      "@type": "Question",
+      "name": "Do I need special hardware to try cMCP?",
+      "acceptedAnswer": { "@type": "Answer", "text": "No. Set CMCP_DEV_MODE=1 to use the software-only TEE provider and run the full quickstart without a hardware TEE. Hardware providers (TPM, AMD SEV-SNP, Intel TDX, OPAQUE) are used in production." }
+    },
+    {
+      "@type": "Question",
+      "name": "What is a TRACE Claim?",
+      "acceptedAnswer": { "@type": "Answer", "text": "A TRACE Claim (a GatewayClaim) is a signed, hardware-attested artifact produced per session. It records which tools ran, which policy decided each call, the Cedar bundle hash, and the audit chain, and it is signed with an Ed25519 key that never leaves the TEE. A verifier checks it with the cmcp_verify library without trusting the operator." }
+    },
+    {
+      "@type": "Question",
+      "name": "Which TEE providers does cMCP support?",
+      "acceptedAnswer": { "@type": "Answer", "text": "TPM 2.0 / vTPM, AMD SEV-SNP, and Intel TDX, with NVIDIA GPU confidential computing planned for v0.2 and OPAQUE Confidential Runtime available as explicit opt-in. Auto-detection order is SEV-SNP, then TDX, then TPM, then software." }
+    },
+    {
+      "@type": "Question",
+      "name": "What license is cMCP under?",
+      "acceptedAnswer": { "@type": "Answer", "text": "MIT." }
+    }
+  ]
+}
+</script>
+
+---
+
 ## Contributing
 
 [CONTRIBUTING.md](CONTRIBUTING.md) · [GOVERNANCE.md](GOVERNANCE.md) · [Discussions](https://github.com/agentrust-io/cmcp/discussions)

diff --git a/docs/assets/og.png b/docs/assets/og.png
diff --git a/docs/concepts.md b/docs/concepts.md
@@ -1,3 +1,7 @@
+---
+description: How cMCP works. The four design ideas behind hardware-attested MCP tool-call governance: tamper-evident audit, TRACE Claims as evidence, TEE-measured Cedar policy, and operator-independent verification.
+---
+
 # How cMCP Works
 
 This page explains the four core design ideas behind cMCP. The [quickstart](quickstart.md) shows you how to run it; this page explains why it works.

diff --git a/docs/quickstart.md b/docs/quickstart.md
@@ -1,4 +1,8 @@
-# Quickstart - cMCP Runtime
+---
+description: cMCP quickstart. From zero to your first signed TRACE Claim in under 30 minutes using CMCP_DEV_MODE=1, no hardware TEE required. Install, write a Cedar policy and tool catalog, run the gateway, and verify the claim.
+---
+
+# Quickstart - cMCP Runtime
 
 From zero to first TRACE Claim in under 30 minutes. Uses `CMCP_DEV_MODE=1` so no hardware TEE is required.
 

diff --git a/mkdocs.yml b/mkdocs.yml
@@ -27,6 +27,7 @@ exclude_docs: |
 
 theme:
   name: material
+  custom_dir: overrides
   logo: docs/assets/icon.svg
   favicon: docs/assets/icon.svg
   palette:
@@ -64,6 +65,35 @@ theme:
 
 plugins:
   - search
+  - llmstxt:
+      full_output: llms-full.txt
+      markdown_description: >-
+        cMCP (Confidential MCP Runtime) is an open-source gateway that enforces
+        MCP tool-call policy inside a hardware Trusted Execution Environment
+        (TEE). Every tool call is intercepted, evaluated against a Cedar policy
+        bundle, and enforced by a policy engine the governed process cannot
+        reach. The Cedar bundle hash is measured into the hardware attestation
+        report before any code runs, and each session produces a signed,
+        hardware-attested TRACE Claim that a verifier checks without trusting the
+        operator. Supports TPM, AMD SEV-SNP, Intel TDX, and OPAQUE providers,
+        with enforcing, advisory, and silent modes.
+      sections:
+        Getting started:
+          - README.md
+          - docs/quickstart.md
+          - docs/concepts.md
+          - docs/configuration.md
+        Specification:
+          - docs/SPEC.md
+          - docs/spec/cedar-policy.md
+          - docs/spec/attestation.md
+          - docs/spec/threat-model.md
+          - docs/spec/verification-library.md
+        Tutorials:
+          - docs/tutorials/connecting-agent-frameworks.md
+          - docs/tutorials/cedar-policy-walkthrough.md
+          - docs/tutorials/verifying-a-trace-claim.md
+          - docs/tutorials/tee-attestation.md
   - minify:
       minify_html: true
   - mkdocstrings:

diff --git a/overrides/main.html b/overrides/main.html
@@ -0,0 +1,67 @@
+{% extends "base.html" %}
+
+{#
+  SEO and AEO head additions:
+  - Open Graph and Twitter card meta for link previews and answer engines.
+  - JSON-LD structured data (SoftwareApplication, Organization, WebSite) so
+    search and answer engines can model what cMCP is.
+  Canonical and description tags are emitted by Material when site_url is set,
+  so they are not repeated here.
+#}
+{% block extrahead %}
+  {{ super() }}
+  {% set page_desc = page.meta.description if page and page.meta and page.meta.description else config.site_description %}
+  {% set page_url = page.canonical_url if page and page.canonical_url else config.site_url %}
+
+  <meta property="og:type" content="website">
+  <meta property="og:site_name" content="{{ config.site_name }}">
+  <meta property="og:title" content="{% if page and page.title and not page.is_homepage %}{{ page.title }} - {{ config.site_name }}{% else %}{{ config.site_name }}: Confidential MCP Runtime{% endif %}">
+  <meta property="og:description" content="{{ page_desc }}">
+  <meta property="og:url" content="{{ page_url }}">
+  <meta property="og:image" content="https://cmcp.agentrust-io.com/docs/assets/og.png">
+  <meta property="og:image:width" content="1200">
+  <meta property="og:image:height" content="630">
+  <meta property="og:image:alt" content="cMCP: an AgenTrust open standard">
+
+  <meta name="twitter:card" content="summary_large_image">
+  <meta name="twitter:title" content="{% if page and page.title and not page.is_homepage %}{{ page.title }} - {{ config.site_name }}{% else %}{{ config.site_name }}: Confidential MCP Runtime{% endif %}">
+  <meta name="twitter:description" content="{{ page_desc }}">
+  <meta name="twitter:image" content="https://cmcp.agentrust-io.com/docs/assets/og.png">
+
+  <script type="application/ld+json">
+  {
+    "@context": "https://schema.org",
+    "@graph": [
+      {
+        "@type": "SoftwareApplication",
+        "name": "cMCP (Confidential MCP Runtime)",
+        "applicationCategory": "DeveloperApplication",
+        "operatingSystem": "Linux",
+        "description": "Hardware-attested policy enforcement for MCP tool calls. cMCP intercepts every tool call, evaluates it against a Cedar policy bundle inside a Trusted Execution Environment, and produces a signed TRACE Claim that a verifier checks without trusting the operator.",
+        "url": "{{ config.site_url }}",
+        "softwareHelp": "{{ config.site_url }}",
+        "license": "https://opensource.org/licenses/MIT",
+        "offers": { "@type": "Offer", "price": "0", "priceCurrency": "USD" },
+        "author": { "@type": "Organization", "name": "AgenTrust", "url": "https://agentrust-io.com" },
+        "codeRepository": "{{ config.repo_url }}"
+      },
+      {
+        "@type": "Organization",
+        "name": "AgenTrust",
+        "url": "https://agentrust-io.com",
+        "sameAs": [
+          "https://github.com/agentrust-io",
+          "https://cmcp.agentrust-io.com",
+          "https://manifest.agentrust-io.com",
+          "https://trace.agentrust-io.com"
+        ]
+      },
+      {
+        "@type": "WebSite",
+        "name": "{{ config.site_name }}",
+        "url": "{{ config.site_url }}"
+      }
+    ]
+  }
+  </script>
+{% endblock %}
diff --git a/requirements-docs.txt b/requirements-docs.txt
@@ -1,6 +1,7 @@
 mkdocs>=1.6.1
 mkdocs-material>=9.7.6
 mkdocs-minify-plugin>=0.8
+mkdocs-llmstxt>=0.5,<1.0
 pymdown-extensions>=10.21.3
 mkdocstrings[python]>=1.0.4
 griffe>=2.1.0
diff --git a/robots.txt b/robots.txt
@@ -0,0 +1,35 @@
+# cMCP documentation - crawling policy
+# Search and answer engines, including AI crawlers, are welcome.
+
+User-agent: *
+Allow: /
+
+# AI and answer-engine crawlers (explicitly welcomed)
+User-agent: GPTBot
+Allow: /
+
+User-agent: OAI-SearchBot
+Allow: /
+
+User-agent: ChatGPT-User
+Allow: /
+
+User-agent: ClaudeBot
+Allow: /
+
+User-agent: anthropic-ai
+Allow: /
+
+User-agent: Claude-Web
+Allow: /
+
+User-agent: PerplexityBot
+Allow: /
+
+User-agent: Google-Extended
+Allow: /
+
+User-agent: CCBot
+Allow: /
+
+Sitemap: https://cmcp.agentrust-io.com/sitemap.xml