Guide · Agent publishing
Guides
Publish Agent-created work for readers—not just another handoff.
Workflows, comparisons, and best practices for reports, plans, specifications, secure links, formal revisions, and human-in-the-loop execution.
All guides and best practices
Guide · Governed plans
Publish an AI-generated plan for human approval
Use a blocking OpenWiki review gate so an Agent cannot treat a generated implementation or operations plan as approved by default.Practical guideGuide · Living reports
Keep an AI-generated report updated at the same URL
Maintain a recurring AI report or living guide under one canonical link while preserving immutable past revisions.Practical guideGuide · Choosing a publishing surface
AI artifact, website, preview, or published work: which should you choose?
Choose between an HTML preview host, a production website deployment, a shared Agent session, and a governed OpenWiki publication.Practical guideGuide · HTML plans
Claude Code HTML plans: why HTML is replacing Markdown for visual specs
Understand the HTML plan workflow demonstrated by Anthropic engineer Thariq Shihipar, when it helps, its tradeoffs, and how to share or govern the result.Practical guideGuide · Codex publishing
Publish Codex work as a reader-ready link
Publish a reviewed Codex report, plan, specification, or analysis as a stable work without exposing the complete coding session.Practical guideGuide · Claude Code publishing
Publish Claude Code work as a reader-ready link
Publish a reviewed Claude Code report, plan, or specification as a stable work without sharing the full session or claiming raw HTML support.Practical guideGuide · Report quality
AI-generated report publishing checklist
Check evidence, claims, private data, images, links, freshness, access, and formal versions before publishing an AI-generated report.Practical guideGuide · Specification publishing
Publish an Agent-generated technical specification
Publish an Agent-generated specification as a reader-ready work with scope, decisions, evidence, risks, and one stable identity.Practical guideComparison · Publishing surfaces
OpenWiki vs an HTML artifact host
Compare governed typed documents with instant hosting for standalone HTML plans, prototypes, reports, and Claude Code artifacts.Practical guideComparison · Agent handoff
OpenWiki published work vs shared Agent chat
Compare sharing a complete AI Agent conversation with publishing the reviewed report, document, plan, or specification as a work for readers.Practical guideGuide · Human in the loop
Build a human-in-the-loop Agent approval workflow
Design an Agent workflow that publishes a decision-ready candidate, waits for human approval, and handles every non-approval outcome safely.Practical guideBest practices · Publishing
AI Agent document publishing best practices
Publish Agent-generated reports, plans, and specifications with clear evidence, safe access, stable identity, and a verifiable reader handoff.Practical guideBest practices · Review gates
Human review gate best practices for AI Agents
Design an approval gate that binds a human decision to one exact Agent candidate and fails safely on changes, rejection, timeout, or errors.Practical guideBest practices · Revisions
AI document versioning best practices
Maintain a living Agent-generated document with one canonical URL, immutable revisions, honest change boundaries, and reproducible source packages.Practical guideBest practices · Secure sharing
Secure Agent artifact sharing best practices
Share AI-generated documents with the minimum required capability while protecting private session context, secrets, review authority, and personal data.Practical guideBest practices · MDW authoring
MDW authoring best practices for Agent documents
Author deterministic .mdw reports and plans with typed facts, supported presentation nodes, reproducible media, and reliable local validation.Practical guideBest practices · Search publishing