PetBox documentation

Public, no sign-in required — safe to open from an agent or share by link.

PetBox overview

What PetBox is, the modules (config, logs, data, dashboard, tasks/memory/sessions) and how to consume them — REST/MCP and the published client libraries. Start here.
Agent onboarding (staged runbook)

The end-to-end sequence — mint on Connect → one petbox-wire command → comprehension → first real plan — with a validation gate at every stage. Start here when bringing an agent online.
Wire a project (petbox-wire CLI)

The operator guide to the CLI: wiring a project, the commands (update / apply / doctor / roles / profile), generated agent files, the offline cache and the exit codes.
MCP surface reference

What PetBox exposes over MCP, the board/node model, the full tool list, and the SKILL.md contents petbox-wire installs. Also the manual registration steps for harnesses it doesn't cover yet (omp, pi).
Methodology cheatsheet (agent)

The operational contract for the spec / work / idea rails: board kinds, node fields, links, effects, the approve gate, and what goes where.
Methodology — the model (philosophy)

Why the rails are shaped this way: the two processes, the entities, the lifecycle, and computed surveyability.