What LazyCodex Is
A thin distribution layer that packages OmO (oh-my-openagent) as the Codex agent harness — project memory, planning, execution, and verified completion, "omo in Codex, packaged for the lazy."
LazyCodex (LZX) packages OmO (oh-my-openagent) as the agent harness for Codex. OmO is the engine that gives Codex discipline agents, parallel orchestration, multi-model routing, skills, hooks, and verified completion. LazyCodex bundles that engine into a repeatable Codex setup you install with one command.
"LazyVim made Neovim usable for the rest of us. LazyCodex does the same for Codex."
Think LazyVim for lazy.nvim, but for Codex. It is omo in Codex, packaged for the lazy.
A thin distribution layer
LazyCodex does not reimplement the agent harness — it distributes one. The core engine is OmO, and npx lazycodex-ai install is shorthand for npx --yes --package oh-my-openagent omo install --platform=codex. What LazyCodex adds is a clean, memorable entry point and a Codex-shaped install.
lazycodex/
├── src/ → oh-my-openagent (OmO submodule — the engine)
└── packages/
└── web/ → the lazycodex.ai documentation siteEverything that does the work — agents, routing, skills, hooks, verification — lives in the OmO submodule. LazyCodex is the layer that points Codex at it.
What the harness gives Codex
Once installed, Codex runs the full OmO harness:
| Capability | What it means in Codex |
|---|---|
| Discipline agents | Sisyphus orchestrates Hephaestus, Oracle, and Librarian — a full AI dev team |
| Parallel execution | Independent subtasks fan out to multiple agents at once |
| Multi-model routing | The right GPT model is chosen automatically per task category |
| Skills system | Specialist judgment (review, cleanup, frontend, LSP, AST-grep, and more) |
| Hooks & lifecycle | Hooks fire around session start, prompt submission, and edit-like tool calls — see Hooks & lifecycle |
| Zero config | Sensible defaults out of the box; override only when you want to |
The four ideas
The harness exists to do four things well, end to end:
Project memory
/init-deep generates hierarchical AGENTS.md context across the repository, so agents land with the right files in view before they edit. Run it when the codebase is too large to explain from memory.
Planning
$ulw-plan "what to build" turns a fuzzy idea into a decision-complete plan written to plans/<slug>.md. It plans only — it never writes product code.
Execution
$start-work [plan-name] executes a plan until every checkbox is done, with durable progress, and prints ORCHESTRATION COMPLETE when the plan is finished.
Verified completion
$ulw-loop "task" keeps moving until the result is verified by evidence rather than a hopeful status update. An Oracle must confirm the work before the loop ends.
Install
npx lazycodex-ai installNo global install — always use npx. For a fully autonomous, no-TUI setup:
npx lazycodex-ai install --no-tui --codex-autonomousExplore the concepts
/init-deep
Generates hierarchical AGENTS.md project memory so future agents start with full context.
The ultrawork loop
How $ulw-loop drives work to Oracle-verified completion — evidence over hopeful status, iteration caps, --completion-promise and --strategy, and its place alongside $ulw-plan and $start-work.