llms.txt

# AgenTrust

> Open tools for governing AI agents: check actions against policy, record decisions, and verify signed evidence.

AgenTrust is the ecosystem at https://agentrust-io.com and the GitHub organization https://github.com/agentrust-io. Use the linked project documentation for normative requirements, release versions, licenses, and deployment prerequisites.

## Start here

- [Overview and architecture](https://agentrust-io.com/): The components, trust boundaries, governance, and evidence links.
- [10-minute tool-call tutorial](https://agentrust-io.com/quickstart/): Write a policy, observe a denied call, and inspect a signed session record on a laptop. Software mode provides no hardware isolation or hardware-backed provenance.
- [Weight Custody Manifest (WCM)](https://wcm.agentrust-io.com/): Bind model-weight identity and custody terms to key-release policy. The local walkthrough uses synthetic evidence and a placeholder key; it does not load a real model or demonstrate hardware protection.
- [Runnable demos](https://agentrust-io.com/demos/): Software examples for policy decisions, evidence verification, delegation, and model-weight custody. Follow each demo's stated prerequisites and limits.
- [Marketplace search](https://agentrust-io.com/marketplace/): Find integrations by framework, technology, and use case.
- [Complete marketplace catalog](https://agentrust-io.com/marketplace/catalog/): Dated integration descriptions and source links, readable without JavaScript or external catalog requests. Listings do not establish endorsement, certification, or hardware validation.

## Technology and trust boundaries

- [Agent Manifest](https://manifest.agentrust-io.com/): Declares agent identity, capabilities, and intended authority. Verification of a declaration does not prove that every runtime action complied with it.
- [Confidential MCP (cMCP)](https://cmcp.agentrust-io.com/): Evaluates routed tool calls against policy and records the decisions. Hardware deployments place the runtime inside a Trusted Execution Environment; confidentiality also depends on egress policy and deployment configuration. The agent and upstream tool server are separate boundaries.
- [Confidential A2A (cA2A)](https://ca2a.agentrust-io.com/): A developer-preview trust profile for Agent2Agent delegation, with attenuated authority, attestation, sealed channels, and provenance. Assurance depends on the evidence and checks performed by the deployment.
- [TRACE](https://trace.agentrust-io.com/): Portable, signed runtime evidence. Hardware provenance requires attestation verification against a trusted root. A signature alone does not establish hardware origin or the truth of all recorded claims.
- [WCM public SDK and demos](https://wcm.agentrust-io.com/): Public evaluation path for weight custody. Key release, encrypted artifact delivery, runtime leases, and key-copy handling are separate responsibilities. The specification is in pre-standardization review; use the public material linked from this page.
- [AgenTrust Telemetry](https://agentrust-io.com/telemetry/): An alpha, backend-neutral event contract and reference SDKs for governance facts, caller-owned OpenTelemetry, and evidence finalization. It is not a policy engine or a proof that recorded events are true.
- [TRACE Registry](https://agentrust-io.com/registry/): Public anchors and signed checkpoints for offline verification. Inclusion, a signer's stated time, independent time evidence, and consistency of log history are different checks. A registry entry does not validate a record's claims.
- [Agentic controls](https://agentrust-io.com/go/): Thirty requirements with stable individual permalinks and OpenCRE mappings. A mapping is not certification or a conformance result.

## Specifications, schemas, and tests

- [TRACE schema](https://agentrust-io.com/schema/trace-v0.2.json): Machine-readable record schema.
- [TRACE conformance suite](https://tests.agentrust-io.com/): Test requirements, error codes, and implementation guidance.
- [cA2A extension identifier](https://agentrust-io.com/extensions/ca2a/v0.1/): Profile discovery and Agent Card declaration.
- [Example cA2A Agent Card](https://agentrust-io.com/examples/ca2a-agent-card.json): Documentation fixture, not a live agent endpoint.
- [GitHub organization](https://github.com/agentrust-io): Source, releases, issues, governance, and contribution instructions.
- [Governance resource directory](https://governance.agentrust-io.com/): Curated tools, standards, and research.

## Community and stewardship

- [Governance and funding](https://agentrust-io.com/#community): Published roles and evidence links. Sponsorship, project ownership, standards hosting, partnership, and adoption are distinct relationships. The NSF submission is in review, not an award.
- [Fellowship status](https://agentrust-io.com/#fellowship): Current application and selection information. Use this page rather than dates copied from older announcements.
- [Architecting at Scale](https://agentrust-io.com/#book): The field guide and its public companion code.

## More documentation

- [TRACE summary](https://trace.agentrust-io.com/llms.txt)
- [Agent Manifest summary](https://manifest.agentrust-io.com/llms.txt)
- [cMCP summary](https://cmcp.agentrust-io.com/llms.txt)
- [cA2A summary](https://ca2a.agentrust-io.com/llms.txt)
- [Sitemap](https://agentrust-io.com/sitemap.xml): Canonical public pages on this host.