Back to Details

using-arise

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Using Arise

使用Arise

Overview

概述

Arise is a local-first orchestration CLI for coding-agent work. Prefer review-first commands before any execution, and treat persisted 
.arise/
state as the record of what happened.
Arise是面向编码Agent工作的本地优先编排CLI。在执行任何操作前请优先使用优先评审类命令，将持久化存储的
.arise/
状态作为所有操作的记录依据。

When to Use

适用场景

  • Asked to use 
    arise
    , 
    .arise/
    , 
    roadmap
    , 
    live
    , 
    history
    , 
    resume
    , 
    replay
    , or 
    recover
    .
  • Need to turn a strict PRD or GitHub issue body into a reviewable YAML plan.
  • Need to inspect saved plans, run state, worktrees, or recovery bundles.
  • Do not use for ordinary repo edits that do not involve Arise workflows.
  • 要求使用
    arise
    、
    .arise/
    、
    roadmap
    、
    live
    、
    history
    、
    resume
    、
    replay
    或
    recover
    的场景
  • 需要将严格的PRD或GitHub issue正文转换为可评审的YAML计划时
  • 需要查看已保存的计划、运行状态、工作树或恢复包时
  • 请勿用于不涉及Arise工作流的普通仓库编辑操作

Quick Reference

快速参考

GoalCommandNotes
Use an installed binary
arise ...
Normal end-user path.
Use a source checkout
cargo run -p arise-cli -- ...
Use when contributing from the Arise repo. The package name is 
arise-cli
; 
cargo run -p arise --
fails in this workspace.
Check readiness
... doctor
Read the selected provider profile and next action before planning.
Initialize config
... init
Safe to rerun.
Plan safely from a PRD
... start --prd <file> --dry-run --no-auto-run
Dry-run still writes 
.arise/plans/
artifacts.
Plan from a GitHub issue
... start --github-issue owner/repo#123 --dry-run --no-auto-run
Use only when the issue body already follows the strict PRD shape.
Review the generated planopen 
.arise/plans/<slug>/generated-plan.arise.yaml
Also inspect 
revisions/rev-<n>/
before execution.
Execute a reviewed plan
... run .arise/plans/<slug>/generated-plan.arise.yaml
Non-dry-run execution needs a Git repo with at least one commit so worktrees can be created.
Inspect run state
... live
, 
... history --limit 10
Start here before rerunning anything.
Recover safely
... recover inspect <run-id>
Use 
resume
for interrupted runs; use 
replay
after fixing an 
integration-blocked
phase; use 
recover repair
only when inspection recommends it.
Work from a roadmap
... roadmap inspect
, 
... roadmap plan
, 
... roadmap sessions
, 
... roadmap resume
Use when the repo already treats 
docs/roadmap.md
as a first-class Arise input.
目标命令说明
使用已安装的二进制文件
arise ...
普通终端用户使用路径
使用源码检出版本
cargo run -p arise-cli -- ...
从Arise仓库贡献代码时使用。包名为
arise-cli
，在此工作区中执行
cargo run -p arise --
会失败
检查就绪状态
... doctor
制定计划前请先读取选定的提供商配置文件和下一步操作指引
初始化配置
... init
可安全重复运行
基于PRD安全生成计划
... start --prd <file> --dry-run --no-auto-run
试运行模式仍会写入
.arise/plans/
工件
基于GitHub issue生成计划
... start --github-issue owner/repo#123 --dry-run --no-auto-run
仅当issue正文完全符合严格PRD格式时使用
评审生成的计划打开 
.arise/plans/<slug>/generated-plan.arise.yaml
执行前也请查看
revisions/rev-<n>/
目录
执行已通过评审的计划
... run .arise/plans/<slug>/generated-plan.arise.yaml
非试运行模式的执行需要至少有一次提交的Git仓库，以便创建工作树
查看运行状态
... live
, 
... history --limit 10
重新运行任何任务前请先执行此操作
安全恢复
... recover inspect <run-id>
中断的运行任务使用
resume
；修复完
集成阻塞
阶段的问题后使用
replay
；仅当检查结果建议时才使用
recover repair
基于路线图工作
... roadmap inspect
, 
... roadmap plan
, 
... roadmap sessions
, 
... roadmap resume
仅当仓库已将
docs/roadmap.md
作为Arise一级输入时使用

Minimal Flow

最简流程

bash
arise doctor
arise start --prd path/to/prd.md --dry-run --no-auto-run
find .arise/plans -maxdepth 2 -name generated-plan.arise.yaml
If the plan is sound, review the saved YAML and only then run 
arise run ...
.
bash
arise doctor
arise start --prd path/to/prd.md --dry-run --no-auto-run
find .arise/plans -maxdepth 2 -name generated-plan.arise.yaml
如果计划合理，请先评审已保存的YAML文件，确认无误后再运行
arise run ...
。

Common Mistakes

常见错误

  • Treating 
    --dry-run
    as side-effect free. Planning still writes 
    .arise/
    artifacts.
  • Skipping plan review and going straight to execution.
  • Using 
    resume
    for a blocked phase checkpoint instead of inspecting first and then using 
    replay
    .
  • Running mutating execution in a repo with no commit, which breaks worktree creation.
  • Ignoring provider readiness. Even dry-run planning still depends on a working planner provider.
  • Sharing 
    .arise/
    bundles without review; they can contain prompts, logs, file paths, and issue content.
  • 误以为
    --dry-run
    没有副作用：生成计划的过程仍会写入
    .arise/
    工件
  • 跳过计划评审直接执行
  • 对阻塞的阶段检查点使用
    resume
    ，而非先检查再使用
    replay
  • 在没有提交记录的仓库中执行变更操作，会导致工作树创建失败
  • 忽略提供商就绪状态：即使是试运行的计划生成步骤也依赖可用的规划器提供商
  • 未评审就分享
    .arise/
    包：其中可能包含提示词、日志、文件路径和issue内容