AI Disclosure
How this specification was made, and by whom.
This file states what in PP-SPEC-079 came from a human and what came from an AI assistant. It is published so that readers, implementers, standards bodies, and examiners can judge the provenance of the work for themselves.
Summary
The method is Craig Ellrod's. The words were drafted with Claude.
The Coal Mine Protocol, PP-SPEC-079, was conceived, designed, and decided by Craig Ellrod, a human. Claude, an AI assistant made by Anthropic, was used as a drafting, structuring, and editing tool under his direction. No part of the protocol's concept or purpose originated with the AI. Structural proposals listed below were drafted by Claude and are pending Craig Ellrod's acceptance as of this 0.1.0 draft; every section is subject to his review before any release is tagged.
Ratio
| Dimension | Human | AI |
|---|---|---|
| Conception and every normative decision | 80% | 20% |
| Design details and naming | 55% | 45% (Claude proposed: Environment Classes, Parity Declaration, Claim Sheet, Comparability Class, the name Coal Mine (after the human rejected Gauntlet), Self-Provided rules, Regulatory Regime axis, Extension Policy table) |
| 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 itemized record of what it proposed and what the author directed, 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, conforming to PP-SPEC-077 Section 21.15. The two forms agree; if they ever differ, the JSON is wrong and must be corrected.
This version (0.1.3, 2026-10-13)
Human: - Decision to remove remaining maintainer powers from 13.6 and 13.7 - All review and acceptance
AI: - Text of the 13.6 and 13.7 changes
Version 0.1.2 (2026-10-13)
Human: - Decision to make Proof Witness independence from the Lab mandatory, remove the named operator, and limit the maintainer to editorial responsibility, in response to the Open Verification Lab leadership question on operator independence - Accepted the Section 14 wording (maintainer editorial only; conformance from the records) - All review and acceptance
AI: - Identified the independence gaps by reviewing 079 against that question - Text of 4.2.3, 4.5, and Section 14, and the matching changelog, FAQ, and conformance entries
Version 0.1.1 (2026-10-12)
Human: - The legacy-model diagnosis: vendor-funded lab reports and vendor-run POCs - In-Situ versus Independent Lab; vendor-run forbidden - Rejecting Gauntlet as overused - Generalizing from vendor products to any control at SUT/DUT/CUT/AUT, any provenance - Applicability to legacy and agentic environments - Asking for a government or reserved class and for extension placement rules - Drift coverage - All review and acceptance
AI: - Text as written - Section structure - Provider and Control Provenance terms - 4.4 Self-Provided Engagements - 5.5 Regulatory Regime and sovereign Registry - 9.5.5 class versus scale - Section 13 Extension Policy
Proposed by the AI, pending the author's explicit acceptance: - Regulatory Regime value list - Extension promotion threshold of three Engagements
AI systems: Anthropic Claude (Fable 5.1, Opus 5.5); OpenAI ChatGPT (banner image only).
Images: - docs/banner.jpg: generated by OpenAI ChatGPT, supplied by Craig Ellrod. Decorative; not part of the specification. - docs/coal-mine.svg: drawn by Claude from the specification, reviewed by Craig Ellrod.
Human: Craig Ellrod
The following originated with Craig Ellrod and reflect his judgment, experience, and decisions:
- The diagnosis: that the two legacy models, the vendor-funded independent lab report and the vendor-run proof of concept, are respectively unreal and self-attested, and that neither is proof. This comes from his own career on both sides of those models, including co-authoring the 2007 WAF Test Methodology and the DUT and SUT competitive test method in Technical Marketing (2008).
- The shift: that the vendor product becomes part of the SUT, DUT, CUT, and AUT, and that its performance is measured by the buyer's infrastructure rather than by the vendor or by a lab the vendor pays.
- The two engagement modes: in the customer's environment (production, pre-production, QA, development), where the product produces its own efficacy telemetry while it operates; and an independent lab such as HACKERverse performing the same motion with no connection to the customer's live environment.
- That a vendor-operated test is not proof of performance.
- The ownership position: sole authorship, Greater Good Commercial License 1.0 (text CC BY 4.0), patent rights reserved, mandatory attribution.
- All review, acceptance, rejection, and revision direction.
AI: Claude (Anthropic)
Claude was used on 2026-10-10 through claude.ai (model Claude Fable 5.1). Its contributions were:
- Drafting the specification text from Craig Ellrod's description of the shift and the two modes, and from the published text of PP-SPEC-077, 078, 013, 006, 025, and 004.
- Proposing the structure: Environment Classes, Parity Declarations, Reference Environments, the Engagement Manifest, the Claim Sheet mapping to Gold canaries, the Right of Reply, the Performance Record and Series, Comparability Classes, and Integration Points IP-M1 to IP-M5. These are drafting proposals pending his review.
- Proposing the name "The Coal Mine Protocol" from his phrase "run the Mine Plan."
- Drafting the README, glossary, FAQ, conformance checklist, model procurement language, schemas, citation metadata, and this disclosure.
Claude did not originate the shift, the two modes, or the prohibition on vendor-run testing, and made no normative decision that Craig Ellrod has not reviewed.
Working method
The work was directed and iterative. Craig Ellrod stated what the protocol must do and why; Claude drafted; Craig Ellrod read, corrected, redirected, renamed, and in several cases rejected drafts outright and restated the requirement. The same-day sequence of versions in the Revision History reflects this process. Where the AI drafted text that Craig Ellrod did not change, that is because he read it and agreed with it, not because it went unread.
What this means for authorship
Craig Ellrod claims authorship of this specification on the basis of its conception, design, every normative decision in it, and the selection, direction, review, and arrangement of all drafted text. AI output was used as a tool in the same way a human technical writer or editor would be used, with the difference that the tool is disclosed here.
Nothing in this file changes the terms in NOTICE.md or LICENSE.md.
Maintenance
This file will be updated with each release. If a future version contains material drafted without AI assistance, or AI contributions of a different kind, this file will say so.
Craig Ellrod Nebulonium, Inc., Proof Protocol 2026-10-10