minimax-ai/mcode-stack

handoff-doc

Write a hand-off document so a fresh session or another person can continue the current task without the chat history.

ソースを見る

使い方

  1. インストールコマンドをコピーし、自分の環境へ導入する前にソースを確認します。
  2. 必要なツール、認証情報、対応環境を原文で確認します。
  3. 入力と成果物を指定し、出力を資料と照合します。
npx skills add https://github.com/minimax-ai/mcode-stack --skill handoff-doc

開始前の確認

  • 導入にはモデル契約、API枠、ファイルへのアクセス許可は含まれません。
  • スター数や導入数は活動の指標であり、独立した品質・安全性試験ではありません。
リポジトリの原文

見出し、例、コード、表、リンク、参照画像を含む原文を表示しています。

Hand-off Doc

Important state should not live only in the context window. A short, readable hand-off doc lets you clear the session, start a new one, or pass the work to someone else, and continue with exactly what you consider important.

Companion to MCode Engineering Notes 01 and 03.

1. Gather from real sources

Before writing, check the actual state instead of relying on memory of the conversation:

  • git status, git log for the branch, and the current diff.
  • The test or build commands that were run and their latest real results.
  • Any existing design, progress or todo files in the repository.
  • Open issues, PRs or review comments linked to the task.

If something cannot be verified, mark it "unverified".

2. Write the doc

Save it in the repository (for example docs/handoff/<task>.md) or wherever the project keeps working notes. Ask before committing it. Use this template:

markdown
# Hand-off: <task>

Updated: <date>   Branch: <branch>   Last commit: <short sha>

## Goal
One or two sentences. What "done" looks like, and the acceptance criteria.

## Decisions made
- <decision> - because <reason>

## Constraints to keep
- <constraint the next session must not break, e.g. "do not change the public API">

## Ruled out
- <approach> - why it was dropped (so nobody tries it again)

## Current state
- Done: ...
- In progress: ...
- Verified by: <command> -> <result>

## Next steps
1. <the very next concrete action>
2. ...

## Watch out for
- Known pitfalls, flaky tests, environment quirks.

## Pointers
- Key files, functions, issues, PRs.

Keep it short. It is a brief, not a transcript. Leave out debugging narratives and tool output that can be reproduced.

3. Hand over

  • Tell the user where the doc is.
  • In the new session, reference the doc first (for example by mentioning its

path) and ask the agent to restate the goal and the next step before it starts.

  • Update the same doc at each milestone instead of creating new ones.

Output

The path of the doc and a three-line summary: goal, current state, next step.

同じリポジトリから

関連する Skills

すべての Skills
minimax-ai
コミュニティ

engineering-code-review

Use when reviewing code, pull requests, patches, CLs, diffs, or proposed implementations for engineering quality, code health, design, functionality, tests, maintainability, specialist risk, or approval risk. TRIGGER on "review this PR", "code review", "LGTM?", "approve?", "is this code/diff/change safe to merge/deploy?", and local diff reviews. DO NOT TRIGGER for PR descriptions, PR splitting, author review-feedback responses, or non-code safety questions unless code review is requested.

1時間の導入数
1
GitHub Stars
685
更新日
11月10日
minimax-ai
コミュニティ

evidence-first-review

Review an agent's finished work by reading the evidence before its summary, naming what each check proves and what it misses, pulling out the decisions it made silently, and trying to falsify the result before accepting it. Use when a MiniMax Code session, subagent or cloud agent reports a task done, before merging its PR or acting on its analysis.

1時間の導入数
1
GitHub Stars
685
更新日
11月10日
minimax-ai
コミュニティ

harness-experiment

Run a harness change as a controlled experiment, with a small end-to-end eval, one change at a time, repeated runs, and looking at flipped tasks, request count, tokens and cache hits, not just the average score. Use before and after changing prompts, tools, skills, subagent layout, compaction or other agent scaffolding, or when asked whether a change helped.

1時間の導入数
1
GitHub Stars
685
更新日
11月10日
minimax-ai
コミュニティ

lifting-loops

Finds loops where the user keeps acting as a manual step for agents (retrying, judging output, approving, carrying context, starting the same task by hand, doing the agent's work) and proposes how to lift the user out: delete a stale rule, change a skill / AGENTS.md / playbook, add a script or metric, add a skill or scheduled task, or advise the user. Keeps a private loop ledger and promotes Claude Code auto-memory into artifacts. With a task already in the session, reflects on that session; in

1時間の導入数
1
GitHub Stars
685
更新日
11月10日