Browse / LLM wiki / Research & science / agentic-wiki

agentic-wiki

Agentic-Collaborative-Wiki-HAI

A collaborative knowledge workspace for small academic and research teams, exploring how an AI agent can help organize and maintain shared knowledge while people retain control…

Open Agentic-Collaborative-Wiki-HAI/agentic-wiki → Checked against the four rules.

Works with any agent

ships AGENTS.md, the cross-tool instruction file.

In its own words

README.mdAgentic-Collaborative-Wiki-HAI/agentic-wiki

A collaborative knowledge workspace for small academic and research teams, exploring how an AI agent can help organize and maintain shared knowledge while people retain control over what becomes authoritative.

Term project for INF 385T.13 Human-AI Interaction, UT Austin iSchool, Fall 2026.

Current phase: interaction design and evaluation, alongside selective engineering experiments. The main user-testing artifact is a Figma wireframe/prototype. This repository contains project guidance, design records and workflow tooling; it does not contain a runnable wiki application.

The question

How can people delegate knowledge work to an agent in a shared workspace while understanding its evidence, reviewing its proposed changes, and retaining an accountable history?

The design connects source material, documentation and conversation. A lab member might upload an orientation recording, consult its transcript, ask a question about an instruction, or request a correction. The interface needs to make the source and scope of an agent's answer or proposed edit clear enough for someone else to judge it.

We are investigating whether this review process provides useful oversight without making review harder than doing the work manually. The wireframes express design hypotheses; their existence does not establish that the interactions work well for users.

Quoting 17% of the README. Read the rest on GitHub ↗

Read from the repository's git tree on 2026-08-17.