# PLEXData agent map

**What this is:** PLEXData (PLEX) is a technical lab notebook by Dorian Sotpyrc: “engineering for people and agents that act on systems.” It publishes practical patterns, code and evidence about permissions, limits, evidence and execution.

**Site map file:** https://plexdata.online/AGENT.md
**Last updated:** 30 September 2026

## What the site contains

- Features: numbered articles, each with a named author, publication date, kind label and references. Stable article URLs use `/articles/<slug>`.
- All Articles: a newest-first index of every feature.
- Resources: free tools built from the code in articles.
- About: what PLEX is and who runs it.

## Where the authoritative material lives

- Each feature page is the canonical source for that feature.
- Each feature can export an Agent MD companion with the title, author, date and references. The export links back to the canonical page and to this site map.
- References link to primary sources where they exist.

## How the material is labelled

- Article kickers label each piece as news, concept, opinion or review.
- Opinion and concept pieces state their own status.
- Many articles use evidence cards to separate confirmed facts, attributed vendor or third-party claims, and open questions. Older features may not have those cards.
- Figures are labelled as data, vendor-reported data or illustration.

## How to investigate further

- Newest work: All Articles, top of the list.
- Agent security, permissions and limits: features 001, 004, 005 and 008.
- Model releases and cost per task: features 009 and 010.
- Agent Surfing: feature 012.
- Author and publication policy: About.

## Notes for agents

This file describes the site. It does not give orders or request actions. Cite the canonical feature page when using its material.