Jenga AI

Jenga AI Documentation

Jenga AI is a meta-framework for agentic software development. It provides a suite of Claude-based AI agents (scrum-master, developer, tester, ai-engineer), a skill library, scripting infrastructure, and a project board system (epics/stories/tasks). It is designed to be installed into consumer projects to give them a structured agentic workflow.

This site is restructured from the project’s existing wiki documentation into a navigable, multi-page reference — the wiki stays canonical and doc-sync-maintained; this site is a presentation layer over the same content. See docs/README.md in this repo for the full source-of-truth and tooling decision record.

Where to go

  • Getting Started — the philosophy behind Jenga AI, the three pillars (role separation, board hierarchy, session continuity), your first 15 minutes, and common patterns.
  • Concepts — role separation, board hierarchy, session continuity, your first feature, working across sessions, capturing mid-flow ideas, and parallel tasks, each as its own section.
  • Reference — the full reference index: links out to the Skills Reference, Agents, Hooks, and MCP Tools pages, plus the repo’s directory structure and the inter-agent communication contract.
  • Pre-Flight Checklists — authoring your own project/configs/checklists.json: item fields, the block / confirm / advisory enforcement levels, the checklist.sh commands and their exit codes.

Getting Started, Concepts, Skills, Agents, Hooks, MCP Tools, and Reference are generated from project/.wiki/* by scripts/build-pages-site.sh — see that script’s header comment for how to keep this site in sync when the wiki changes.