Skills
Specialist playbooks LazyCodex layers on top of Commands.
LazyCodex installs the OmO plugin's Skills into Codex. Commands ($ulw-plan, $start-work, $ulw-loop, $init-deep) shape the workflow; Skills provide the specialist judgment used inside that workflow.
Skills load inside a Codex session. The user can explicitly ask for a Skill, and Codex can also follow a playbook when the task matches the Skill description.
Documented Skills
| Skill | Use it for |
|---|---|
review-work | Multi-angle post-implementation review |
remove-ai-slops | Behavior-preserving cleanup of AI-looking code |
debugging | Systematic root-cause debugging |
frontend | Designed UI work instead of generic layout filling |
programming | Strict TypeScript, Rust, Python, and Go discipline |
refactor | Test-verified refactoring, including structural search and rewrite with AST-grep |
LSP | Diagnostics, definitions, references, symbols, and rename |
rules | Project instructions from AGENTS, rules, and instruction files |
comment-checker | Feedback after edit-like work |
Other Installed Skills
As of 4.12.1, the plugin also includes the Skills below. This docs site does not have a page for each yet, but they can load inside a Codex session.
| Skill | Use it for |
|---|---|
ast-grep | AST-aware structural search and rewrite |
git-master | Commits, history investigation, rebase/squash/fixup git work |
lsp-setup | Language server installation and configuration |
lcx-doctor | LazyCodex/Codex install health diagnosis |
lcx-report-bug | LazyCodex/Codex bug issue drafting |
lcx-contribute-bug-fix | Verified bug-fix contribution workflow |
teammode | Coordinating several Codex threads as a team |
ultraresearch | Explicitly requested deep research |
visual-qa | Screenshot and visual verification after UI/TUI changes |
Commands themselves are also installed as Skills, but the user-facing docs collect them separately under Commands.
Where Skills Fit
- Commands own the flow: planning, execution, and verified completion.
- Skills add specialist judgment around that flow: review, cleanup, UI work, language discipline, and code intelligence.
- Skills load only for relevant work, then drop away, keeping the agent prompt lighter.
Skill Docs
review-work
Five agents review completed work in parallel, and all must pass.
remove-ai-slops
Lock behavior with tests, remove AI slop, then verify.
frontend
Design, UI, UX, performance, and visual QA in one Skill.
debugging
Systematic root-cause debugging.
refactor
Test-verified refactoring with structural search and rewrite.
Commands
The
$ulw-* and $init-deep workflows that Skills run around.