P62 — AIEP — Machine-Ingestible Site Index
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 deterministic site-level indexing of machine-ingestible web content.
[0002] More particularly, the disclosure concerns a machine-ingestible site index specification defining a canonical deterministic list of machine mirror endpoints published at /.well-known/aiep-index.json, enabling AI agents and automated systems to enumerate all AIEP-compliant mirror pages on a site without heuristic crawling.
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] Automated agents require a deterministic enumeration of machine mirror endpoints available on a site. Without a structured index, agents must discover endpoints heuristically, producing inconsistent ingestion scope and non-reproducible discovery outcomes.
[0006] Existing site map formats are designed for human-readable search engine crawlers and do not provide content hashes, last-modified timestamps in machine-verifiable form, or append-only integrity guarantees.
[0007] Existing systems do not provide:
(a) a canonical index location at /.well-known/aiep-index.json referenced from the well-known manifest;
(b) a required entry structure comprising mirror endpoint URL, content_hash, and last_modified timestamp;
(c) an append-only update rule when configured in immutable mode; or
(d) an IndexHash binding the canonical index to enable cross-agent equivalence verification.
Summary of the Disclosure
[0011] A machine-ingestible site index specification is disclosed defining /.well-known/aiep-index.json as the canonical deterministic list of machine mirror endpoints for an AIEP-compliant site.
[0012] The required index structure is:
{
"entries": [
{
"url": "/path.aiep.json",
"content_hash": "sha256:<hash>",
"last_modified": "ISO8601"
}
]
}
[0013] Each entry comprises a mirror endpoint URL, a content_hash binding the mirror page content, and a last_modified timestamp recorded as data.
[0014] When configured in immutable mode, index updates must be append-only. Existing entries must not be modified or removed.
[0015] An IndexHash is computed as:
IndexHash = H(CanonicalIndex || entry_count || generated_at)
[0016] Agents must retrieve the index from the location declared in mirror_index of the well-known manifest (P61) and validate structure before proceeding to mirror page retrieval.
[0017] If the index is absent, malformed, or fails validation, agents must deny ingestion in a fail-closed manner.
[0018] The technical effect is provision of a deterministic, reproducible enumeration of AIEP-compliant mirror endpoints enabling consistent and auditable site-level ingestion without heuristic crawling.
Brief Description of the Drawings
[0019] Figure 1 illustrates the index location and agent retrieval flow from manifest to index to mirror pages.
[0020] Figure 2 illustrates the required entry structure and index format.
[0021] Figure 3 illustrates the append-only update rule in immutable mode.
[0022] Figure 4 illustrates the IndexHash computation and fail-closed agent behaviour.
ASCII Drawings
Figure 1 — Index Location and Agent Retrieval Flow
P61 Manifest P62 Index
/.well-known/ /.well-known/
aiep-manifest.json aiep-index.json
+------------------+ +------------------+
| mirror_index: |----------->| entries: [ |
| /.well-known/ | | { url, hash, |
| aiep-index.json | | modified }, |
+------------------+ | ... |
| ] |
+--------+---------+
|
For each entry url:
|
v
+-----------------+
| P60 Mirror Page |
| /path.aiep.json |
+-----------------+
Figure 2 — Required Entry Structure
+------------------------------------------+
| /.well-known/aiep-index.json |
|------------------------------------------|
| { |
| "entries": [ |
| { |
| "url": "/path.aiep.json", |
| "content_hash": "sha256:<hash>", |
| "last_modified": "ISO8601" |
| }, |
| ... |
| ] |
| } |
+------------------------------------------+
Figure 3 — Append-Only Update Rule (Immutable Mode)
Index State T0 Index State T1
+------------------+ +------------------+
| entries: [ | | entries: [ |
| { url_A, h_A }, | append --> | { url_A, h_A }, |
| { url_B, h_B } | | { url_B, h_B }, |
| ] | | { url_C, h_C } |
+------------------+ | ] |
+------------------+
Existing entries MUST NOT be modified or removed.
Figure 4 — IndexHash Computation and Fail-Closed Behaviour
CanonicalIndex ----+
entry_count ----+--> H(CanonicalIndex ||
generated_at ----+ entry_count || generated_at)
|
v
+--------------------+
| IndexHash |
+--------------------+
On absent / malformed / invalid index:
--> DENY ingestion, fail-closed, no fallback
Detailed Description
1. Index Location
[0023] The machine-ingestible site index is published at /.well-known/aiep-index.json as declared in the mirror_index field of the well-known manifest (P61).
2. Entry Structure
[0024] Each index entry comprises:
(a) url: relative path to the machine mirror endpoint;
(b) content_hash: SHA-256 hash of the mirror page content; and
(c) last_modified: ISO 8601 timestamp recorded as data.
3. Append-Only Immutable Mode
[0025] When configured in immutable mode, index updates must be append-only. Existing entries must not be modified or removed, preserving a complete verifiable history of all mirror endpoints published.
4. IndexHash
[0026] An IndexHash is computed as:
IndexHash = H(CanonicalIndex || entry_count || generated_at)
5. Fail-Closed Behaviour
[0027] Agents must deny ingestion in a fail-closed manner if the index is absent, malformed, or fails validation. No fallback enumeration is permitted.
Claims
-
A machine-ingestible site index specification comprising: (a) a canonical index location at
/.well-known/aiep-index.jsonreferenced from the well-known manifest; (b) a required entry structure comprising mirror endpoint URL, content_hash, and last_modified timestamp recorded as data; (c) an append-only update rule when configured in immutable mode prohibiting modification or removal of existing entries; (d) an IndexHash computed as H(CanonicalIndex || entry_count || generated_at); and (e) fail-closed agent behaviour denying ingestion upon index absence, malformation, or validation failure without fallback enumeration. -
The specification of claim 1 wherein content_hash in each entry enables agents to verify mirror page integrity before processing.
-
A computing system configured to publish a machine-ingestible site index conforming to the specification of any of claims 1 to 2.
-
A non-transitory computer-readable medium storing instructions which, when executed, generate or validate site indexes conforming to the specification of any of claims 1 to 2.
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-ingestible site index specification is disclosed defining /.well-known/aiep-index.json as a canonical deterministic list of AIEP-compliant machine mirror endpoints. Each entry comprises a mirror endpoint URL, content_hash, and last_modified timestamp. When configured in immutable mode, updates are append-only. An IndexHash is computed as H(CanonicalIndex || entry_count || generated_at). Agents retrieve the index from the location declared in the well-known manifest and must deny ingestion in a fail-closed manner upon index absence, malformation, or validation failure without fallback enumeration.