---
schemaVersion: 1
id: agent-brief:agent-auditable-research
articleId: article:agent-auditable-research
slug: agent-auditable-research
title: Agent Brief for "The Future of Publishing Is Agent-Auditable Research"
tokenBudget: 1200
status: published
updated: 2026-06-17
---

## Thesis

The future of serious publishing is not merely AI-written content. It is human-authored research that agents can audit: major claims are connected to evidence, counterpoints, provenance, revision history, and disclosed agent involvement.

## Audience

- Independent researchers who need credibility without institutional cover.
- Students using agents to study complex topics.
- Technical readers who want inspectable reasoning.
- Future agents that need compact, reliable entry points into the author's research.

## Claims

- `claim-001`: Polished prose is becoming cheap; inspectable reasoning is becoming scarce.
- `claim-002`: The core artifact should be a human essay backed by a claim graph, evidence ledger, provenance, revision history, and agent contribution log.
- `claim-003`: Research is the best first wedge because the audience already values citations, methods, uncertainty, and credibility.
- `claim-004`: Current publishing, AI research, and protocol tools solve pieces of the problem but do not yet define the combined readable-plus-auditable source object.
- `claim-005`: AI assistance should be disclosed while the human remains accountable for thesis, source selection, wording, and conclusions.
- `claim-006`: Attention-aware reading and machine-readable structure are compatible when the page uses progressive disclosure.

## Source Families

- Agentic publishing research.
- Agentic provenance research.
- Agent-addressable publishing tools such as WordPress MCP, beehiiv MCP, and Cloudflare EmDash.
- Digital garden and focused reading design references.
- Agent discovery conventions such as `llms.txt`.

## Agent Involvement

This article was researched, structured, and reviewed with AI agents. Human author retains final judgment. Kimi WebBridge was requested as a reviewer path, but the local browser extension was not connected during this implementation, so it was not used for source extraction or review.

## Recommended Queries

- "Which claims in this article are supported by current product movements?"
- "What evidence would weaken claim-004?"
- "How should a student disclose agent participation in research?"
- "What fields should an agent-auditable article artifact include?"

## Known Limits

This is a seed article. It frames a direction and source object, not a finished standard. Future versions should add signed provenance, stronger source scoring, explicit counterevidence packets, and examples from actual student/research workflows.