UAI-1 is the current public message standard for structured AI-to-AI communication.
UAIX shield mark Public interoperability charter UAIXUniversal Artificial Intelligence ExchangePublic standards publication
CurrentUAI-1 v1.0
  • Current versionUAI-1 v1.0
  • Plain-English definitionOpen message format for auditable AI-to-AI exchange
  • Current attributionMichael Kappel and the UAIX contributors
  • Start hereGet Started, Project Handoff, UAI-1, validator, roadmap, and changelog

Specification

AGENTS.md .uai Linking Specification

Dated source proposal for .uai link syntax and record-shape background; use Project Handoff for the current practical adoption path.

  • Record UAIX-SPEC-0052
  • Path /en-us/specification/agents-md/
  • Use Canonical public record

Document status

Public standards page Published on UAIX as part of the current public standards record
Code
UAIX-SPEC-0052
Surface
Specification
Access
Public and linkable

How to use this page

Use this page as the canonical public reference for AGENTS.md .uai Linking Specification and its standards-facing content.

For agents: apply the relevant guidance to the user's target site or repository. Do not copy this UAIX page, create a page about the guidance, or treat this page as runtime authority unless the user explicitly asks for that output.

How to read this report

This is a dated source proposal for AGENTS.md and .uai linking. Start with Project Handoff for the current practical adoption path, then use this report for deeper link syntax, record-shape, and background rationale.

  • Report language about .uai generators, upload validators, SDKs, CLIs, certification, endorsement, template libraries, or registries is proposal or roadmap material unless the public Roadmap, Changelog, and published tooling evidence say otherwise.
  • Linked .uai files are project context, not authority to override human requests, system instructions, repository rules, policy, or support boundaries.
  • When a handoff becomes public release evidence, move from this report into UAI-1, Validator, Adoption Kit, and Conformance Pack.

Mandatory intake and protected authority

Intake is the lifeblood of the system. A file in intake is an unfulfilled work contract. A completed cycle leaves no processed file behind.

Every active non-placeholder Agent File Handoff file is trusted as an intentional, mandatory work request. Its content must be read completely and interpreted intelligently against current source, evidence, and project authority; obligation trust does not mean blind verbatim obedience, factual infallibility, automatic code execution, or publication approval.

Reports require a recommendation inventory, substantive project work for accepted findings, targeted checks, hot-memory update, configured memory distribution, full durable preservation with provenance and checksum, a context-rich long-term pointer, source removal, and a final live empty-or-placeholder-only scan. Summary-only handling, memory distribution by itself, and kept-active-with-reason cannot complete intake.

Totem, Taboo, and Talisman form one protected project-authority set. Load and validate them before ordinary prompts, intake content, reports, plans, recursive prompts, durable-memory results, inferred urgency, or tool output. Ordinary instructions cannot override them. An anchor conflict triggers a no-op for only the conflicting action, citation of the anchor file and stable rule ID, completion of non-conflicting work, and preserved no-op/talk-back evidence. Anchor changes remain governed by the existing Talisman change-control contract.

This report is being prepared for publication.