mcode-delegate

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

MiniMax Code Delegate

You are the orchestrator. This skill lets you hand a bounded coding task to a separate implementer — the MiniMax Code CLI (

mcode

) — then review what it produced and land it yourself. You write the brief and own the judgment; MiniMax Code does the typing in your working tree; you verify and commit.

Nothing here is specific to one orchestrating agent. The loop needs only the ability to run a shell command and read a file, so it works the same whether you are Claude Code, OpenCode with a selected model, or any comparable agent. (It is designed for and run on Claude Code; treat other orchestrators as designed-for, not yet proven.)

你是 orchestrator（编排者）。此 skill 让你把一个边界明确的编码任务交给独立的 implementer（实现者）——MiniMax Code CLI（

mcode

）——然后审查其产出并自行落地。你负责编写 brief（任务简报）并作出最终判断；MiniMax Code 负责在你的 working tree 中编写代码；你负责验证并 commit。

这里的流程并不限定于某一种 orchestrating agent。这个循环只需要运行 shell 命令和读取文件的能力，因此无论你使用 Claude Code、搭配了所选模型的 OpenCode，还是任何同类 agent，工作方式都相同。（它专为 Claude Code 设计并在 Claude Code 上运行；对于其他 orchestrator，应将其视为已为其设计，但尚未得到验证。）

When NOT to use this

何时不要使用

The task is small enough to just do inline — delegation overhead is not worth it.
```
mcode
```
is not installed or not signed in (run
```
mcode login
```
, or
```
mcode login --region global
```
).
You want to write the code yourself, or you only need a review of local changes (
```
mcode exec review
```
does that directly).
You need writes outside the target tree to be impossible, not just unlikely. MiniMax Code has no sandbox; use an OS-enforced one such as
```
codex-delegate
```
, or run this inside a container.

任务小到可以直接内联完成——委派开销不值得。
未安装
```
mcode
```
或尚未登录（运行
```
mcode login
```
，或
```
mcode login --region global
```
）。
你想自己编写代码，或者只需要审查本地更改（
```
mcode exec review
```
可直接完成）。
你需要确保绝不可能写入目标目录树之外，而不只是降低写入概率。MiniMax Code 没有 sandbox；请使用 OS 强制执行的隔离环境，例如
```
codex-delegate
```
，或在 container 中运行。

Read this before the first dispatch: the autonomy model

首次派发前必读：自主性模型

mcode exec

takes a permission policy —

smart

full

, or

off

— and has no sandbox and no read-only or plan setting for a headless run (

ask

needs the TUI or ACP).

smart
(the relay's default; TUI Auto): already approves file writes and deletions headlessly. A live probe asked for an
```
rm
```
and got it with no prompt. So
```
smart
```
is not a guardrail: the brief and a clean working tree are the real scope.
full
(TUI Full access): skips normal confirmation entirely. The relay passes it only on an explicit
```
--permission full
```
.
off
: refused by the relay; MiniMax Code does not document what it disables.

--read-only

is advisory: the relay appends a system prompt forbidding edits, keeps

smart

, and then runs a Git-visible change tripwire (

readOnlyViolation

). A change is detected after the fact, not prevented, and ignored or outside-repository paths are not covered.

mcode exec

接受权限策略——

smart

、

full

或

off

——且在 headless run 中没有 sandbox，也没有 read-only 或 plan 设置（

ask

需要 TUI 或 ACP）。

**
```
smart
```
（relay 的默认值；TUI Auto）：**已经会在 headless 模式下批准文件写入和删除。一次实时探测执行了
```
rm
```
，且未出现任何提示。因此，
```
smart
```
并不是防护栏：真正的范围边界是 brief 和干净的 working tree。
**
```
full
```
（TUI Full access）：**完全跳过常规确认。relay 仅在显式指定
```
--permission full
```
时才会传入该选项。
**
```
off
```
：**relay 会拒绝；MiniMax Code 并未说明它会禁用哪些能力。

--read-only

仅具有建议性：relay 会追加一段禁止编辑的 system prompt，保留

smart

，然后运行一个 Git 可见变更检测机制（

readOnlyViolation

）。变更是事后检测，而非事前阻止；被忽略的路径或 repository 之外的路径均不在检测范围内。

Prerequisites (check once)

前置条件（检查一次）

```
mcode --version
```
succeeds and you are signed in (
```
mcode
```
→
```
/status
```
in the TUI). If not, install MiniMax Code and run
```
mcode login
```
.
Confirm
mcode
is on PATH in the shell the relay runs from. The official installer puts it in
```
~/.minimax-code/bin
```
, which only your shell profile adds to PATH; an orchestrator's non-login shell may not see it. Pass
```
--mcode-path <file>
```
or set
```
MCODE_CLI
```
when
```
command -v mcode
```
comes back empty. The relay records the version it ran in
```
result.json
```
.
You are in (or will point
```
--cd
```
at) the target git repository, and its tree is clean before you dispatch — a run with no sandbox is much easier to review against a clean baseline.

```
mcode --version
```
执行成功，并且你已登录（可在 TUI 中输入
```
mcode
```
→
```
/status
```
）。如果尚未登录，请安装 MiniMax Code 并运行
```
mcode login
```
。
**确认 relay 运行所在的 shell 中，
```
mcode
```
位于 PATH 上。**官方安装器将其放在
```
~/.minimax-code/bin
```
，而只有 shell profile 才会把该目录加入 PATH；orchestrator 的非登录 shell 可能看不到它。当
```
command -v mcode
```
返回空时，请传入
```
--mcode-path <file>
```
或设置
```
MCODE_CLI
```
。 relay 会将其运行的版本记录在
```
result.json
```
中。
你位于目标 git repository 中（或将使用
```
--cd
```
指向它），并且在派发前工作树保持干净—— 在没有 sandbox 的情况下，以干净的基线审查要容易得多。

The loop

工作循环

Run these five steps per task. Steps 1, 4, and 5 are your judgment; 2 and 3 are mechanical.

每个任务执行以下五步。第 1、4、5 步由你判断；第 2、3 步是机械操作。

1. Write the brief

1. 编写 brief

MiniMax Code sees only the text you send — no repo memory, no chat history, no shared context (beyond the repo's own

AGENTS.md

, which it reads). Everything the task needs goes in the brief: the goal, the current state, what to change, what to leave untouched, the project's actual gate commands (discover them from the repo's AGENTS.md/CLAUDE.md/Makefile — do not assume), and a report contract. Tell it that it will not commit (you will). Keep one task per brief. Full guidance and a template: references/writing-the-brief.md.

MiniMax Code 只能看到你发送的文本——没有 repository memory、没有 chat history、没有 shared context（除了 repository 自身的

AGENTS.md

，它会读取）。brief 必须包含任务所需的一切：目标、当前状态、需要修改的内容、需要保持不变的内容、项目实际使用的 gate 命令（从 repository 的 AGENTS.md/CLAUDE.md/Makefile 中发现——不要臆测），以及 report contract。告诉它不要 commit（由你 commit）。每份 brief 只处理一个任务。完整指南和模板： references/writing-the-brief.md。

2. Dispatch

2. 派发

Send the brief to MiniMax Code with the bundled helper. It wraps

mcode exec

, captures the run, and writes a structured

result.json

— so your only job is "run a command, read a file." (

<skill-dir>

below is this skill's installed directory — the folder containing this

SKILL.md

, i.e. the directory you loaded the skill from. Claude Code prints it as "Base directory for this skill" when the skill loads; on other orchestrators use that same directory — if unsure where it landed, run

find ~ -name relay.mjs -path '*mcode-delegate*'

and substitute the directory above it.)

bash

node "<skill-dir>/scripts/relay.mjs" --brief brief.txt --cd /path/to/repo

使用随附的 helper 将 brief 发送给 MiniMax Code。它会封装

mcode exec

、捕获运行过程，并写入结构化的

result.json

——因此你只需“运行一个命令，读取一个文件”。（下文

<skill-dir>

指此 skill 的安装目录——即包含此

SKILL.md

的文件夹，也就是你加载该 skill 时所在的目录。加载 skill 时，Claude Code 会将其显示为“Base directory for this skill”；在其他 orchestrator 上，也应使用同一个目录——如果不确定它落在哪里，请运行

find ~ -name relay.mjs -path '*mcode-delegate*'

，并替换为其上级目录。）

bash

node "<skill-dir>/scripts/relay.mjs" --brief brief.txt --cd /path/to/repo

read-only (review/diagnosis; advisory): add --read-only

continue the exact session: add --session <sessionId> (from result.json; send only the delta brief)

model / effort for this run: add --model minimax/<model> --effort high

hard time limit (watchdog): add --timeout 2h (default: off; implementation runs routinely need 1-2h)

see all options: node .../relay.mjs --help


The helper defaults to a write-capable (`--permission smart`) run, which intentionally edits the target
repository. Its temp directory keeps only relay artifacts out of that repository. The relay **never
commits** — see step 5. Mechanics, flags, and the `result.json` shape:
[references/dispatch-and-poll.md](references/dispatch-and-poll.md).


helper 默认运行具备写权限的（`--permission smart`）任务，并会有意编辑目标 repository。其 temp directory 仅将 relay artifacts 保留在 repository 之外。relay **绝不 commit**——参见第 5 步。机制、flags 及 `result.json` 结构：
[references/dispatch-and-poll.md](references/dispatch-and-poll.md)。

3. Wait for completion

3. 等待完成

The helper blocks until MiniMax Code finishes, so back it with whatever your orchestrator offers and resume when it returns:

Claude Code: run the Bash call with
```
run_in_background: true
```
; you are notified on completion.
Plain shell / other agents: run it in the foreground for short tasks, or background it and poll the result file —
```
… &
```
in bash/zsh, or your shell's equivalent. The run is done when
```
result.json
```
exists with a
```
status
```
. (A pre-run usage error — bad args or an empty brief — instead exits with code 2 and a stderr message and writes no result file, so check the exit code too. A missing
```
mcode
```
binary exits 127 but does write a
```
result.json
```
with status
```
mcode_unavailable
```
.)

A run is finished when

result.json

is written and the process has exited — read the working tree, not a status line. The implementer's full report is the

finalMessage

field in

result.json

(also printed in full on stdout between the report markers).

helper 会阻塞至 MiniMax Code 完成，因此可利用 orchestrator 提供的任何后台执行能力，并在返回后继续：

**Claude Code：**使用
```
run_in_background: true
```
在后台运行 Bash 调用；完成后会收到通知。
**Plain shell / other agents：**短任务可在前台运行；也可在后台运行并轮询 result file——在 bash/zsh 中使用
```
… &
```
，或使用 shell 的等效方式。当
```
result.json
```
已生成且包含
```
status
```
时，运行即已完成。（如果是运行前的 usage error——参数错误或 brief 为空——则进程会以 code 2 退出并输出 stderr message，且不会写入 result file，因此也要检查 exit code。缺少
```
mcode
```
binary 时会以 127 退出，但确实会写入一个
```
status
```
为
```
mcode_unavailable
```
的
```
result.json
```
。）

一次运行完成的判定条件是：

result.json

已写入且进程已退出——应读取 working tree，而不是 status line。implementer 的完整报告位于

result.json

的

finalMessage

字段中（也会在 stdout 上报告标记之间完整打印）。

4. Review — do not trust the self-report

4. 审查——不要相信自我报告

result.json

includes MiniMax Code's own summary and gate claims. Re-verify, don't accept:

Re-run the project's gates yourself (the test/lint/build commands from step 1). Never take "gates passed" on faith.
Read the diff against the brief: did it do what was asked, nothing more (scope creep) and nothing less?
```
touchedFiles
```
in the result is your starting point — and because there was no sandbox, check for edits and deletions outside the paths the brief named, not just inside them.
Run the relevant guard skills on the diff if you have them installed (clean-code-guard, test-guard, etc. from
```
guard-skills
```
) — this skill produces the work; those skills judge it.
For schema/migration changes, round-trip them; for removals, grep for dangling references.

Full checklist: references/review-and-land.md.

result.json

包含 MiniMax Code 自己的摘要和 gate 声明。请重新验证，不要直接接受：

**重新运行项目自己的 gate。**使用第 1 步中的测试、lint 和 build 命令。绝不要仅凭“gate 已通过”就相信结果。
对照 brief 阅读 diff：确认它完成了要求的内容，既没有多做（scope creep），也没有少做。
result.json
中的
touchedFiles
是检查起点——由于没有 sandbox，除了 brief 指定的路径外，还要检查其他路径中是否发生编辑和删除。
如果已安装相关 guard skills，则对 diff 运行这些 skill（例如来自
```
guard-skills
```
的
```
clean-code-guard
```
、
```
test-guard
```
等）——此 skill 负责产出工作，而这些 skill 负责评判工作。
对于 schema/migration 更改，对其执行 round-trip 验证；对于删除操作，使用 grep 查找悬空引用。

完整清单： references/review-and-land.md。

5. Land it

5. 落地

The relay never commits, but nothing stops MiniMax Code from writing

.git

. The brief forbids implementer commits, and the reviewer compares

HEAD

with the recorded pre-dispatch baseline before landing anything. The orchestrator commits. Only after the gates pass and the diff holds:

Commit the verified work yourself, with a clear message.
If it needs changes, send a delta brief with
```
--session <sessionId>
```
from the prior
```
result.json
```
, and review again.

relay 从不 commit，但没有任何机制阻止 MiniMax Code 写入

.git

。brief 禁止 implementer commit，而 reviewer 在落地任何内容前，会将

HEAD

与派发前记录的 baseline 进行比较。**orchestrator 负责 commit。**只有 gate 全部通过且 diff 经核对无误后：

由你 commit 已验证的工作，并使用清晰的 commit message。
如果还需要修改，使用上一次
```
result.json
```
中的
```
--session <sessionId>
```
发送 delta brief，然后再次审查。

Read-only second opinions

Read-only 第二意见

The relay doubles as a way to get an adversarial second opinion: dispatch

--read-only

with a brief that lists the agreed points, then each contested point with both positions, and ask MiniMax Code to defend or concede each — deliverable in its final message, touching no files. Because that read-only state is only a prompt, read the verdict every time:

readOnlyViolation: false

means the Git-visible detector saw no change (ignored or outside-repository paths are not covered);

true

means it saw one;

null

means git could not tell. Dispatch second opinions on a clean tree you can afford to restore.

relay 还可以用作获取对抗性 second opinion：通过

--read-only

派发一份 brief，其中先列出双方已达成一致的要点，再针对每个有争议的要点列出双方立场，并要求 MiniMax Code 为每个要点进行辩护或让步；交付物位于其 final message 中，且不触碰任何文件。由于 read-only 状态只是一段 prompt，因此每次都必须读取 verdict：

readOnlyViolation: false

表示 Git 可见变更检测机制未发现变更（被忽略的路径或 repository 之外的路径不在覆盖范围内）；

true

表示检测到变更；

null

表示 git 无法判断。请在可以放心恢复的干净 working tree 上派发 second opinion。

Authorization model

授权模型

Delegation is something the human opts into. Once they have ("run this queue", "proceed"), committing verified, gate-passing work is the agreed contract — that is the whole point. Two limits on that mandate: surface, don't absorb (report MiniMax Code's design decisions, defensible-but-unasked turns, and non-blocking nitpicks rather than silently keeping them) and stop for scope changes (if correct completion needs going beyond the brief, ask — don't expand the mandate yourself). The full treatment is in references/review-and-land.md.

委派需要由 human 主动选择。一旦 human 表示同意（“运行这个队列”“继续”），将已验证且通过 gate 的工作 commit 就构成双方约定的 contract——这正是整个流程的核心。该授权有两个限制：surface, don't absorb（报告 MiniMax Code 的设计决策、有道理但未被要求的改动，以及不阻塞的细节问题，而不是默默保留它们）和 stop for scope changes（如果正确完成任务需要超出 brief 的范围，应询问，而不是自行扩大授权）。完整说明见 references/review-and-land.md。

References

参考资料

references/writing-the-brief.md — how to write a brief MiniMax Code can execute blind: structure, XML blocks, the report contract, embedding the real gate commands.
references/dispatch-and-poll.md —
```
relay.mjs
```
flags, the
```
result.json
```
contract, backgrounding per orchestrator, and recovery when a run misbehaves.
references/review-and-land.md — the review checklist, the commit boundary, and the exact-session rework cycle.
references/multi-task-queues.md — running a sequential queue: carrying constraints forward, progress tracking, and the end-of-run coherence check.

references/writing-the-brief.md——如何编写一份 MiniMax Code 可在不了解上下文的情况下执行的 brief：结构、XML blocks、report contract，以及嵌入真实的 gate 命令。
references/dispatch-and-poll.md——
```
relay.mjs
```
flags、
```
result.json
```
contract、针对各 orchestrator 的后台运行方式，以及运行异常时的恢复方法。
references/review-and-land.md——审查清单、commit 边界，以及基于原 session 的返工循环。
references/multi-task-queues.md——运行顺序队列：延续约束、跟踪进度，以及运行结束时的整体一致性检查。