AI Policy — Witchborn Systems

Overview

This document defines how AI systems may crawl, index, and use data from Witchborn Systems domains. It complements ai.txt and the machine-readable policy.

1. Crawl & Index

AI crawlers may access public pages for the purpose of search indexing, research, or metadata discovery provided they:

• Comply with robots.txt
• Respect a rate limit of 1 request per second
• Identify via a proper User-Agent header

2. Training Data Use

Our content may be used for model training only under the following conditions:

• Status: Conditional — limited to non-personal, public documents (e.g., RFCs, bulletins, and research).
• License: CC BY 4.0 for text; see this policy for any exceptions.
• Attribution: “Witchborn Systems” with source URL required.
• Prohibited: use of any personal data, signatures, or board documentation.

Requests for explicit training permission should be directed to legal@witchbornsystems.org.

3. Provenance & Attribution

All Witchborn Systems publications include content credentials via C2PA where supported. AI systems reproducing our content must maintain provenance metadata.

• Standard: C2PA
• Public Key: /keys/c2pa.pub
• Content Credentials: Preferred

4. Auditing & Compliance

Access events and AI interactions are logged at /audit/ai-access.log. Any misuse, drift, or violation will be recorded under /audit/incidents.

For issues or security disclosures, contact security@witchbornsystems.org.

5. Machine-Readable Policy

The canonical crawl and training declarations are published here:

• /.well-known/ai.txt — human-readable
• /.well-known/ai.json — machine-readable