AI-285 student project

C=US

A national person and AI-agent registry using X.500 with LDAP access, distinct from the existing Federal Bridge PKI.

Demo presentation: this site and its maple-quality workflow are conceptual and locally tested. They are not a production authorization service or a claim of confirmed organizational participation.

Use cases Documents

Color illustration of a dragon within an equal-sign motif — The first line of Hexagram 1, Qian / The Creative, is commonly rendered as Hidden dragon. Do not act. - 潛龍，勿用 (*Qián lóng, wù yòng*). Translation wording assisted by Google Gemini. See Chinasage and Selfgazer.

What have you learned from this website?

Nodes light up once you've actually visited them in this browser (tracked locally, not sent anywhere) — a map of where you've been, not just a list of links.

1White House accord scorecard → 2China/US compatibility → 3Governance models evaluation → 4Hardware signing scorecard → 5Break/fix correction cycle → 6Utah SEDI walkthrough → 7Blind bias check → 8Signed attestation log (raw)

Use cases

Maple syrup grading Utah tart cherry grading Anxi Tieguanyin oolong (China) AI consultants Agent access disputes

All use cases

Explore the project

Hands-on

Explore agent registration lab Explore agent authority lab Verify XAdES-T signatures by agents

Background reading

Documents: proposed architecture poster Technical glossary State-endorsed identity concept Technology / culture stack YubiKey evaluation guide Agentic risks case study Encryption and standards FAQ Identity vs. runtime: an agentic control plane Control-plane architectures vs. real compliance standards NewAgentic moral values: Claude’s constitution, faith leaders, the Vatican and the world values map NewThe Singularity: joining, or listing? Life’s major transitions and why the directory keeps agents and people apart Newp(doom) feedback loop: published estimates, what the architecture does and does not address, and your own score NewDuty of loyalty, enforced in code: nine scenarios for the SEDI conference NewCEQUS Systemic Risk Control Plane: agent dependency mapping for underwriting, added after the LSE & Swiss Re Institute report NewPost-quantum testing: ML-DSA identity and hybrid ML-KEM results from the local gateway Scorecard: the White House AI accord vs. this project's risks Scorecard: the accord and c=US vs. Chinese AI regulation Scorecard: what a hardware-backed signature actually buys you Claude's own signed evaluation: three governance models, one rubric Correcting a certified AI output: a real break/fix cycle Implementing c=US's sponsor gap with Utah's real SEDI law Recently updatedBlind replication: checking a Jev score for its own bias

YubiKey mTLS registration workflow

A local demonstration of how a WebAuthn security-key ceremony can gate issuance of a short-lived client certificate for an AI agent. The workflow illustrates earned, credential-bound trust; it is not a Yubico certification or a production identity service.

1 · Verify the personRegister and authenticate a cross-platform WebAuthn credential. Evaluator mode requests user verification (PIN or equivalent) plus touch.

2 · Bind the sessionThe server verifies the challenge, origin, relying-party ID, signature counter, and credential used for the authenticated session.

3 · Register the agentThe authenticated session gates creation of a workload identity and issuance of a demo mTLS client certificate.

4 · Present to a serviceThe sample relying service requires a client certificate signed by the demo CA before accepting the TLS connection.

Evaluation boundary: attestation is intentionally none, so the demo does not prove manufacturer or model. Production work still requires agent-generated keys/CSR, protected CA custody, registry authorization, revocation, persistent state, and compliance review. Read the evaluator guide or review the demo source.