Skip to content

uze-sh/uze

unversioned · 617abf717f25Apache-2.0

UZE's own agentic Skills: making a project's instructions context (AGENTS.md) portable, writing the rules that apply only where they are relevant, working in the checkout UZE placed you in, writing the diagrams it keeps under workspace.artifacts, authoring plugins of your own, and the extensions they carry, and trying the marketplace's example plugins — each one reasoning over a deterministic uze CLI that is the only thing that ever mutates or judges anything.

architect

Writing and changing the Mermaid diagrams a project keeps under workspace.artifacts — choosing which diagram and which level a change belongs in, naming it so it reads where a reader expects, and proving it still draws with uze agent artifacts check, which needs no workspace. Use when adding or editing a .mmd/.mermaid file, when asked to diagram or document an architecture, when a diagram does not draw, or when setting up where a project keeps its architecture.

author

Guides creating a uze plugin end to end — marketplace create or select, plugin scaffold, capability flags, check before install, install from the linked marketplace, iterate, publish by pushing — and writing what it carries, a declared extension included (its extension.kdl, its .uze components of a view and a script, a program, a resident script, grants, tests, consent). Use when someone asks to create, author or scaffold a plugin, set up a marketplace of plugins, add an extension to a plugin, write or debug an extension's script, program, source, action or key, when uze agent plugin check refuses an extension, or when one draws "needs consent", "needs enabling in this project" or "<program> not found".

init

Makes this project's instructions context portable across Claude Code, Codex, OpenCode, and Antigravity CLI, using AGENTS.md as the shared baseline. Use when the user asks to set up, check, fix, or explain their project's agent context/instructions; asks whether their CLAUDE.md/GEMINI.md/AGENTS.md is portable between tools; wants an AGENTS.md created or reviewed; mentions switching between coding agents and losing context; or invokes /uze:init or $uze:init directly.

rules

Writing a project's rules, the instructions that apply only when they are relevant, under .agents/rules/. Covers deciding whether something is a rule, AGENTS.md or a Skill, choosing the trigger, writing patterns in the supported glob dialect, wording a rule as a convention, deciding when a rule belongs in a shared plugin and moving it there, and proving it with uze agent rules check. Use when asked to create, add or write a rule, when an instruction "only applies to src/…" or to one kind of file, when AGENTS.md has grown sections that only matter under one path and should be split, or when a rule does not seem to fire.

try

Sets up a project to try uze's workspace extensions in — gsd, focus, contributions — so the person sees what extensions do in the workspace without wiring anything of their own. Use when the person asks to try uze, to see example extensions or plugins, to set up a demo or playground project, or invokes /uze:try or $uze:try directly.

worktree

For an agent uze workspace launched, and no other. Works inside the checkout the workspace placed you in — knowing where you are, committing on your own branch, giving parallel subagents checkouts of their own, and handing your work to UZE's delivery instead of integrating it yourself. Use when coordinating more than one writing agent, when resuming work in an existing checkout, when UZE reports a paused rebase or failed checks on your branch, or when a conflict or an unknown checkout owner needs resolving.