AI Disclosure
How this specification was made, and by whom.
This file states what in PP-SPEC-083 came from a human and what came from an AI assistant, as PP-SPEC-077 Section 21.15 requires.
Summary
What must have an identity, the single arena identity, and what the specification must answer are Craig Ellrod's. The principles, mechanisms, most of the names, and the words were proposed and drafted by Claude, at his direction, and await his review.
Ratio
| Dimension | Human | AI |
|---|---|---|
| Conception and every normative decision | 75% | 25% |
| Design details and naming | 40% | 60% (mean of design 50/50 and naming 30/70; Craig Ellrod set the scope and the single arena identity, Claude proposed the principles, record mechanics, and the names listed below as pending acceptance) |
| Words, code, and schemas as written | 5% | 95% |
| Review, direction, acceptance, decision to publish | 100% | 0% |
Method: self-estimated. Shares drafted by the AI from its record of what the author directed and what the AI proposed, for the author to confirm or correct before release. MAY be re-measured with a tool such as ProofJack.com.
Machine-readable form: ai_disclosure.json. The two forms agree; if they ever differ, the JSON is wrong and must be corrected.
This version (0.1.0, 2026-10-30)
Human (Craig Ellrod): - The requirement that everything in a run has an identity: the arena, the ingress point, the egress point, the traffic pipe, each interface, every agent, and every target - The idea that the entire arena can be identified with one authN and authZ identity - The questions the specification answers: who owns each identity, where it is stored, how it is accessed, and the schema and fields pulled and pushed - Raising identity, authentication, and authorization as missing from the proof machine (2026-10-29) - The rule that reproof spins up entirely new infrastructure rather than reusing the original run's instances (2026-10-29) - The original title Identity 101, the number PP-SPEC-083, the direction to rename it so the name is unique to proof, acceptance of the name ProofID 101 (2026-10-30), and the decision to publish it (2026-10-30) - Review and acceptance (pending)
AI (Claude Opus 5.5), at the author's direction: - The four-layer analysis of identity in the proof machine (operator, machine, bound subject, agent identity and authorization, plus issuer identity), 2026-10-29 - The principles: testers sign and subjects are fingerprinted; nothing vouches for its own independence; identity is not authority - Independent Roots outside the Arena Identity; keys under IP-10 custody outside the environment under test; Scope bound to Integration Points; Delegated Credentials bound to the Declared Policy; asserted versus captured Subject identity - The Identity Record fields and the Identity Manifest, the pull and push sequence, lifecycle, and the split of Subject identity from instance identifiers for reproof - Section structure, all text as written, the two schemas and their examples, and the supporting files
Proposed by the AI, accepted by Craig Ellrod on 2026-10-30: - The name ProofID 101 (0.1.1), chosen from four options
Proposed by the AI, pending the author's explicit acceptance: - The names Keyed Identity, Fingerprinted Identity, Root Identity, Independent Root, Component Identity, Identity Record, Identity Manifest, Delegated Credential, Scope, Arena Operator - The identity kinds and their field names (Section 5, Section 12) - The finding kinds scope_violation and identity_mismatch - The IDN-1 domain tag
AI systems: Anthropic Claude (Opus 5.5).
Nothing in this file changes the terms in NOTICE.md or LICENSE.md.