P60 — AIEP — Machine Mirror Page Standard
Publication Date: 2026-02-26 Status: Open Source Prior Art Disclosure Licence: Apache License 2.0 Author/Organisation: Phatfella Ltd Schema: AIEP_OS_SPEC_TEMPLATE v1.0.1 — https://aiep.dev/schemas/aiep-os-spec-template/v1.0.1
Field of the Invention
[0001] The disclosure relates to machine-readable web content standards within computing environments.
[0002] More particularly, the disclosure concerns a canonical structure for machine-ingestible mirror pages associated with human-readable web pages, enabling deterministic hash-bound evidence ingestion by AI agents and automated systems operating under an Architected Instruction and Evidence Protocol (AIEP) framework.
Framework Context
[0003] This invention operates within an Architected Instruction and Evidence Protocol (AIEP) environment as defined in United Kingdom patent application number GB2519711.2, filed 20 November 2025, the entire contents of which are incorporated herein by reference.
[0004] The present invention extends deterministic canonicalisation, governance, and execution integrity mechanisms defined in the AIEP environment while remaining independently implementable as described herein.
Background
[0005] Web pages are designed for human readability and contain presentational markup, navigation elements, and rendering metadata that is irrelevant to automated evidence ingestion.
[0006] AI agents and automated systems ingesting web content for evidence purposes must process human-oriented markup to extract structured data, introducing non-deterministic parsing variability across implementations.
[0007] Existing machine-readable formats do not provide a standardised deterministic envelope structure binding canonical content to a cryptographic hash, version identifier, and structured evidence fields suitable for AIEP-compliant admissibility evaluation.
[0008] Existing systems do not provide:
(a) a canonical URL convention associating a machine mirror endpoint with each human-readable page; (b) a required envelope structure comprising schema identifier, schema version, content hash, generation timestamp, source URL, structured data, evidence array, probability metadata, and plausibility metadata; (c) a requirement that JSON be canonicalised before hashing; (d) a content hash binding exclusively the data object; or (e) a prohibition on additional facts in the mirror not derivable from the human source.
[0009] There exists a need for a machine mirror page standard that defines a deterministic canonical envelope structure, a URL convention, canonicalisation-before-hashing requirements, and a no-additional-facts constraint.
Summary of the Disclosure
[0011] A machine mirror page standard is disclosed defining a canonical structure for machine-ingestible mirror pages associated with human-readable web pages.
[0012] A machine mirror page is published at a canonical mirror endpoint derived from the human page URL by appending .aiep.json to the path or by publishing at /m2m/{path}.json.
[0013] The required envelope structure comprises:
{
"aiep_version": "1.0.0",
"schema_id": "aiep.web.page.v1",
"schema_version": "1.0.0",
"content_hash": "sha256:<hash>",
"generated_at": "ISO8601",
"source_url": "https://example.com/path",
"data": {},
"evidence": [],
"probability": {},
"plausibility": {}
}
[0014] JSON must be canonicalised before hashing using the rules defined in P63 — AIEP Content Hash Binding Protocol: UTF-8 encoding, sorted keys, no insignificant whitespace, deterministic number formatting.
[0015] The content_hash field must bind the data object only, computed over the canonicalised serialisation of the data object.
[0016] A MirrorPageHash is computed as:
MirrorPageHash = H(CanonicalData || schema_id || schema_version || source_url)
[0017] No additional facts may be present in the mirror page that are not derivable from the human-readable source page at the declared source_url.
[0018] If canonicalisation fails or content_hash cannot be verified, the mirror page is non-admissible and must not be processed.
[0019] The technical effect is modification of web publishing and AI ingestion system behaviour by providing a deterministic canonical evidence envelope enabling reproducible hash-bound machine ingestion of web content.
Brief Description of the Drawings
[0020] Figure 1 illustrates the URL convention mapping human page paths to machine mirror endpoints.
[0021] Figure 2 illustrates the required envelope structure and field relationships.
[0022] Figure 3 illustrates the canonicalisation-before-hashing pipeline for content_hash computation.
[0023] Figure 4 illustrates the no-additional-facts constraint and fail-closed non-admissibility determination.
ASCII Drawings
Figure 1 — URL Convention
Human-Readable Page Machine Mirror Endpoint
+---------------------+ +---------------------------+
| https://example.com | | https://example.com |
| /path | --> | /path.aiep.json |
+---------------------+ +---------------------------+
OR
+---------------------+ +---------------------------+
| https://example.com | | https://example.com |
| /path | --> | /m2m/path.json |
+---------------------+ +---------------------------+
Figure 2 — Required Envelope Structure
+------------------------------------------+
| Machine Mirror Page |
|------------------------------------------|
| aiep_version : "1.0.0" |
| schema_id : "aiep.web.page.v1" |
| schema_version : "1.0.0" |
| content_hash : "sha256:<hash>" |
| (binds data obj only) |
| generated_at : ISO8601 timestamp |
| source_url : canonical source URL |
| data : {} (structured content)|
| evidence : [] (evidence array) |
| probability : {} (probability meta) |
| plausibility : {} (constraint results)|
+------------------------------------------+
| No additional facts not derivable from |
| human source permitted |
+------------------------------------------+
Figure 3 — Canonicalisation-Before-Hashing Pipeline
data object (raw)
|
v
+------------------+
| Canonicalise |
| - UTF-8 encoding |
| - Sort JSON keys |
| - Remove insig. |
| whitespace |
| - Det. number |
| formatting |
+--------+---------+
|
v
+--------+---------+
| SHA-256 Hash |
| content_hash = |
| H(canonical_data)|
+------------------+
Figure 4 — No-Additional-Facts Constraint and Admissibility Gate
+---------------------------+
| Mirror Page Received |
+------------+--------------+
|
v
+------------+--------------+
| Verify content_hash |
| against canonical data |
+------------+--------------+
|
+------+------+
Verified Not Verified
| |
v v
+------------------+ +-----------+
| No-additional- | | NON- |
| facts check | | ADMISSIBLE|
| All facts in | | Fail- |
| mirror derivable | | Closed |
| from source_url? | +-----------+
+--------+---------+
|
+------+------+
Yes No
| |
v v
+----------+ +-----------+
| ADMISSIBLE| | NON- |
| | | ADMISSIBLE|
+----------+ +-----------+
Detailed Description
1. URL Convention
[0024] A machine mirror page is published at a canonical endpoint derived from the human-readable page URL by one of two conventions:
(a) appending .aiep.json to the page path, producing /path.aiep.json; or
(b) publishing at /m2m/{path}.json.
[0025] The URL convention enables automated agents to deterministically locate the machine mirror for any human-readable page without prior configuration.
2. Required Envelope Fields
[0026] The machine mirror page envelope comprises the following required fields:
(a) aiep_version: AIEP protocol version string;
(b) schema_id: canonical schema identifier;
(c) schema_version: schema version string;
(d) content_hash: SHA-256 hash of the canonicalised data object only;
(e) generated_at: ISO 8601 timestamp recorded as data;
(f) source_url: canonical URL of the associated human-readable source page;
(g) data: structured content object;
(h) evidence: array of evidence references;
(i) probability: probability metadata envelope as defined in P66; and
(j) plausibility: plausibility constraint declaration as defined in P67.
3. Canonicalisation Before Hashing
[0027] JSON must be canonicalised before hashing using: UTF-8 encoding, alphabetically sorted keys at all levels, no insignificant whitespace, and deterministic number formatting.
[0028] The content_hash field binds the data object only and is computed over the canonical serialisation of the data object.
[0029] A MirrorPageHash is computed as:
MirrorPageHash = H(CanonicalData || schema_id || schema_version || source_url)
4. No-Additional-Facts Constraint
[0030] No additional facts may be present in the mirror page that are not derivable from the human-readable source page at the declared source_url.
[0031] Agents detecting facts in the mirror page not present in or derivable from the source page must treat the mirror as non-admissible.
5. Fail-Closed Non-Admissibility
[0032] If canonicalisation fails, content_hash cannot be verified, or the no-additional-facts constraint is violated, the mirror page is non-admissible and must not be processed for evidence purposes.
Claims
-
A machine mirror page standard defining a deterministic canonical structure for machine-ingestible mirror pages, the standard comprising: (a) a URL convention associating a machine mirror endpoint with each human-readable page by appending
.aiep.jsonto the path or publishing at/m2m/{path}.json; (b) a required envelope structure comprising aiep_version, schema_id, schema_version, content_hash, generated_at, source_url, data, evidence, probability, and plausibility fields; (c) a requirement that JSON be canonicalised before hashing using UTF-8 encoding, sorted keys, no insignificant whitespace, and deterministic number formatting; (d) a content_hash binding the canonicalised data object only, computed as H(CanonicalData); (e) a MirrorPageHash computed as H(CanonicalData || schema_id || schema_version || source_url); and (f) a prohibition on additional facts in the mirror page not derivable from the human-readable source, with fail-closed non-admissibility determination upon violation. -
The standard of claim 1 wherein content_hash is computed using SHA-256 by default with future algorithm versions declared in the manifest.
-
The standard of claim 1 wherein agents detecting hash verification failure or no-additional-facts violation treat the mirror page as non-admissible without fallback processing.
-
A computing system configured to publish machine mirror pages conforming to the standard of any of claims 1 to 3.
-
A non-transitory computer-readable medium storing instructions which, when executed, cause one or more processors to generate or validate machine mirror pages conforming to the standard of any of claims 1 to 3.
Licence
Any person is granted a perpetual, irrevocable, worldwide, royalty-free licence to make, use, implement, modify, or distribute any system or method described in this disclosure for any purpose, without restriction, under the Apache License 2.0.
A copy of the Apache License 2.0 is available at https://www.apache.org/licenses/LICENSE-2.0
Abstract
A machine mirror page standard is disclosed defining a canonical deterministic structure for machine-ingestible mirror pages associated with human-readable web pages. Machine mirrors are published at canonical endpoints derived from human page URLs by appending .aiep.json or at /m2m/{path}.json. A required envelope structure comprises aiep_version, schema_id, schema_version, content_hash, generated_at, source_url, data, evidence, probability, and plausibility fields. JSON is canonicalised before hashing using UTF-8, sorted keys, no insignificant whitespace, and deterministic number formatting. A content_hash binds the canonicalised data object only. A MirrorPageHash is computed as H(CanonicalData || schema_id || schema_version || source_url). No additional facts not derivable from the human source are permitted. Mirror pages failing hash verification or the no-additional-facts constraint are non-admissible.