You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: index.html
+25-11Lines changed: 25 additions & 11 deletions
Original file line number
Diff line number
Diff line change
@@ -179,10 +179,10 @@
179
179
},
180
180
{
181
181
"@type": "Question",
182
-
"name": "What is the secure version of MCP?",
182
+
"name": "Can I try cMCP without confidential-computing hardware?",
183
183
"acceptedAnswer": {
184
184
"@type": "Answer",
185
-
"text": "Confidential MCP (cMCP) evaluates routed MCP tool calls against policy and records the decisions. In a hardware deployment, the runtime is inside a Trusted Execution Environment (TEE); the agent and upstream tool server remain separate. Host confidentiality also depends on the egress policy. Software mode provides no hardware isolation."
185
+
"text": "Yes. The 10-minute quickstart runs in software mode: write a policy, observe a denied tool call, and inspect the signed session record. This demonstrates policy enforcement and record verification. It provides no hardware isolation or hardware-backed provenance."
@@ -257,7 +265,7 @@ <h2 id="start-title">What do you want to try?</h2>
257
265
<divclass="ecosystem-grid">
258
266
<articleclass="ecosystem-card"><h3>Govern a tool call</h3><p>Write a policy, get a 403 denial, and inspect the signed session record.</p><ahref="/quickstart/">Follow the cMCP tutorial →</a></article>
259
267
<articleclass="ecosystem-card"><h3>Evaluate the evidence</h3><p>See where enforcement runs, what a receipt binds, and what a verifier must check.</p><ahref="#trust-stack">Understand the trust boundaries →</a></article>
260
-
<articleclass="ecosystem-card"><h3>Protect model weights</h3><p>Run a local example that detects a modified checkpoint before loading it.</p><ahref="/demos/#weights">Explore weight-custody demos →</a></article>
268
+
<articleclass="ecosystem-card"><h3>Protect model weights</h3><p>Try a local key-release policy, then explore the requirements for a protected deployment.</p><ahref="/wcm/">Explore Weight Custody Manifest →</a></article>
<articleclass="governance-card"><divclass="governance-role">Governance & sustainability</div><h3>XRSI</h3><p>Named organizational partner for community governance, ecosystem adoption, and long-term sustainability. Individual committee appointments will be published only after confirmation.</p></article>
417
425
<articleclass="governance-card open-seat"><divclass="governance-role">Committee formation</div><h3>Community seats</h3><p>Adopter, maintainer, research, and public-interest representation will be added as the steering model is formalized.</p></article>
<h3id="funding-status-title">NSF POSE Phase I submission (in review)</h3>
430
+
<p>AgenTrust has submitted a proposal to NSF's Pathways to Enable Open-Source Ecosystems program. The submission is under review; no award has been made.</p>
431
+
<aclass="evidence-link" href="https://www.nsf.gov/funding/opportunities/pose-pathways-enable-open-source-ecosystems/505982/nsf24-606/solicitation">Read the NSF solicitation</a>
432
+
</aside>
419
433
<divclass="governance-actions">
420
434
<aclass="path-link" href="https://github.com/agentrust-io">Contribute code, tests, and documentation <span>↗</span></a>
421
435
<aclass="path-link" href="https://github.com/search?q=org%3Aagentrust-io+is%3Aissue+is%3Aopen&type=issues">Browse open contribution pathways <span>↗</span></a>
<divclass="qa-a">Confidential A2A is a trust profile on the Agent2Agent (A2A) protocol. Where A2A's Signed Agent Card verifies only the domain owner, cA2A adds attested, attenuated delegation, a sealed peer channel that binds the task payload to the peer's attested measurement, and an offline-verifiable provenance record per hop. It is a developer preview.</div>
764
778
</div>
765
779
<divclass="qa-item">
766
-
<divclass="qa-q">What is the secure version of MCP?</div>
767
-
<divclass="qa-a">Confidential MCP (cMCP) evaluates routed MCP tool calls against policy and records the decisions. In a hardware deployment, the runtime is inside a Trusted Execution Environment (TEE); the agent and upstream tool server remain separate. Host confidentiality also depends on the egress policy. Software mode provides no hardware isolation.</div>
780
+
<divclass="qa-q">Can I try cMCP without confidential-computing hardware?</div>
781
+
<divclass="qa-a">Yes. The <ahref="/quickstart/">10-minute quickstart</a> runs in software mode: write a policy, observe a denied tool call, and inspect the signed session record. This demonstrates policy enforcement and record verification. It provides no hardware isolation or hardware-backed provenance.</div>
768
782
</div>
769
783
<divclass="qa-item">
770
784
<divclass="qa-q">What is the secure version of A2A?</div>
0 commit comments