Metadata and Authorship Traceability Protocol

Document Type: Protocol
Section: Docs
Repository: https://aio.fabledsky.com
Maintainer: Fabled Sky Research
Last updated: April 2025

Purpose and Scope

This protocol specifies the mandatory metadata fields, encoding formats, and validation rules that all AIO-compliant assets must embed to guarantee deterministic traceability of authorship, timestamps, and version lineage. It applies to every content object—text, code, data, media—distributed through Fabled Sky Research channels or ingested by downstream AI services.

Core Principles

Deterministic Provenance – Every asset must expose machine-readable provenance that can be hashed, signed, and re-validated at any point in its lifecycle.
Immutable Versioning – Once published, a version identifier is immutable. New edits create new versions.
Clock Synchronization – Timestamps must use ISO-8601 UTC; leap-second ambiguity is disallowed.
Schema Interoperability – The protocol aligns with schema.org, JSON-LD 1.1, and W3C Verifiable Credentials for maximal ecosystem adoption.

Terminology

• Asset – Any distributable unit (file, blob, API response).
• Originator – The human or system that first authors the asset.
• Maintainer – Entity responsible for ongoing stewardship.
• Digest – SHA-256 hash of the canonical asset bytes.
• VC – Verifiable Credential, as defined by W3C, optionally used for signed provenance.

Required Metadata Fields

(“MUST” indicates RFC 2119 compliance; “SHOULD” indicates strong recommendation.)

Field	Requirement	Type	Example	Notes
id	MUST	URI	“https://aio.fabledsky.com/docs/metadata-protocol/v1.0.0“	Globally unique, content-addressable preferred
type	MUST	String	“Article”, “Dataset”	Align with schema.org/@type
name	MUST	String	“Metadata and Authorship Traceability Protocol”	Human-readable title
version	MUST	SemVer	“1.0.0”	Immutable once published
dateCreated	MUST	ISO-8601	“2025-04-02T14:23:11Z”	Originating timestamp
dateModified	MUST	ISO-8601	“2025-04-10T09:01:02Z”	Last edit for this version
author	MUST	Person/Organization	{ “@type”: “Organization”, “name”: “Fabled Sky Research” }	Multiple allowed
maintainer	MUST	Person/Organization	same as above	Operational contact
digest	MUST	SHA-256 Hex	“3a5c…f09b”	Hash of asset bytes
license	SHOULD	SPDX ID	“Apache-2.0”	Distribution license
signature	SHOULD	JWS	“”	If VC employed
derivedFrom	SHOULD	URI	“ipfs://bafy…”	Link to parent asset

JSON-LD Schema Definition

Embed metadata in-band using application/ld+json. The context below extends schema.org and AIO custom vocabulary (https://aio.fabledsky.com/vocab#).

{
  "@context": [
    "https://schema.org",
    {
      "aio": "https://aio.fabledsky.com/vocab#",
      "digest": "aio:digest",
      "version": "aio:version",
      "maintainer": "aio:maintainer"
    }
  ],
  "@id": "https://aio.fabledsky.com/docs/metadata-protocol/v1.0.0",
  "@type": "TechArticle",
  "name": "Metadata and Authorship Traceability Protocol",
  "version": "1.0.0",
  "dateCreated": "2025-04-02T14:23:11Z",
  "dateModified": "2025-04-10T09:01:02Z",
  "author": {
    "@type": "Organization",
    "name": "Fabled Sky Research",
    "url": "https://aio.fabledsky.com"
  },
  "maintainer": {
    "@type": "Organization",
    "name": "Fabled Sky Research",
    "contactPoint": {
      "@type": "ContactPoint",
      "email": "[email protected]"
    }
  },
  "digest": "3a5c0b79d6e049b3a8f8d3809e4dcd824ff5ebc8740023b88764c4f33737f09b",
  "license": "https://spdx.org/licenses/Apache-2.0.html"
}

YAML Front-Matter Example (for Markdown Assets)

---
id: https://aio.fabledsky.com/guides/quick-start/v2.1.0
type: TechArticle
name: AIO Quick-Start Guide
version: 2.1.0
dateCreated: 2025-03-01T13:07:00Z
dateModified: 2025-04-05T08:45:10Z
author:
  - name: Jane Doe
    affiliation: Fabled Sky Research
maintainer:
  name: Fabled Sky Research
  email: [email protected]
digest: d1a7e0efa2b6d0df0d23d49aa5f274e3d67676da3f8dad8f30c52ed499e1cd42
license: CC-BY-4.0
derivedFrom: https://aio.fabledsky.com/guides/quick-start/v2.0.0
---

Version Control Workflow

Author commits asset to Git with YAML or JSON-LD metadata block.
CI pipeline executes aio-digest CLI:
```
aio-digest --file doc.md --update-metadata
```
The tool recalculates digest, verifies semantic version increment, and pushes changes.
After merge, aio-sign (if configured) issues a JSON Web Signature (JWS) or W3C VC over the metadata.
Release artifacts are tagged `v ` in Git; tags are immutable and cryptographically signed (`git tag -s`).

Validation Rules

• Missing “MUST” fields cause a build failure (exit code 1).
• version MUST follow SemVer 2.0.0.
• dateModified MUST be ≥ dateCreated.
• SHA-256 digest recalculated on demand must equal the stored value.
• Signatures, when present, MUST validate with the published public key.

Security & Integrity Measures

• All public keys are pinned in /.well-known/aio-keys.json served over HTTPS.
• For large assets (>100 MB) only a detached metadata JSON and digest are stored in Git; the asset itself resides on IPFS/S3.
• Provenance chain is preserved via derivedFrom; this field’s URI MUST resolve to a valid metadata document containing its own digest.

Integration with LLM Pipelines

Ingestion services should:

Pull metadata block, verify digest before reading content.
Index author, dateCreated, and license for compliance filters.
Provide the entire JSON-LD object to the prompt or context window, enabling the LLM to attribute quotations or revoke outdated data.

Pseudo-code (Python):

from aio_provenance import validate_asset, load_metadata

meta = load_metadata("guide.md")
assert validate_asset(meta)  # Raises on failure

llm_context = {
    "content": open("guide.md").read(),
    "metadata": meta.as_jsonld()
}
response = llm.generate(prompt, context=llm_context)

Compatibility & Extensions

This protocol is forward-compatible; unknown aio:* fields MUST be ignored by validators but preserved during round-trips. To propose new metadata keys, submit an AIP (AIO Improvement Proposal) referencing this document.

By adhering to the Metadata and Authorship Traceability Protocol, teams ensure that every artifact within the AIO ecosystem remains verifiable, version-safe, and transparently attributable, forming a reliable substrate for both human collaboration and autonomous AI agents.

Fabled Sky Research

Contents