nhi.sgit.ai / pki

PKI: the cryptographic half of the identity gap

This section is the staging ground for pki.sgit.ai — a site collecting the published work on public key infrastructure, and a registry of agent keys designed from a documented failure. It lives here first because it is the cryptographic half of this site's thesis: the shared-drives research found that no available system gives an agent its own identity, agent-level attribution, or encryption to a specific agent. Per-agent keys in a registry are what would fill those empty rows.

Source. This section is built from the pki.sgit.ai strategy brief (v0.33.59, 16 Aug 2026), captured verbatim with all citations and readable on its document page. The historical claims are externally verifiable, so the participant discipline applies lightly here — the design claims are ours and are stated as checkable rules.

The pages

The history

Why good public key repositories don't exist

The global keyserver network was destroyed in 2019 by an attack its own maintainer called unsalvageable — and the cause was a stated design goal, not a bug. The three abused properties, and what the replacement gave up.

Read →
The design

The four rules, published before the registry exists

Only the owner writes to their own record; revocation is a signed append; records are size-bounded; every entry is signed. Plus: identity vs. mandate, and what vaults do and don't supply.

Read →

Why attempt this now

The registry idea is old; what changed is that vaults supply the layer a key server had to build itself: distribution (a published vault is readable from static hosting and mirrorable by anybody), custody without access (untrusted mirrors are safe by construction), and versioning (the state of a key at any past date is answerable natively). What vaults do not supply — the ownership rule, the size bound, signature checking — is the registry logic, and it is the part to build carefully. The registry becomes a product of existing infrastructure rather than a new system.

The material exists, in depth

729 files in the corpus mention the subject, with dedicated documents going back to February 2026: an architecture debrief, a messaging implementation, model-integration research, a document identity brief, operating modes, a provenance treatment, and a brand strategy. So pki.sgit.ai is a publication and curation problem, like this site — organise by the question a reader arrives with, promote the strongest artefacts to pages. The curation queue is tracked on comms.

Build order (from the brief)

  1. The collection, organised by question, with the February–March architecture material promoted.
  2. The failure page — short, useful, independent of anything being built. Done, here.
  3. The four rules, published as the registry's stated design before the registry exists. Done, here.
  4. A private registry, for our own agents — the smaller problem, with a real user.
  5. Mandate statements, separate from identity statements, revocable independently.
  6. A public registry, last: it inherits every hard problem of the private one, plus trust roots and abuse.

The load-bearing ordering: a registry with one organisation's agents in it is testable; a global one is a commitment.

How this connects to the rest of the site