Skip to content

rumoii/rumo-engineering-skills / skills/rumo-handoff

rumo-handoff

Verified

Use when the user asks to wrap up ("收尾"), hand off, pause, or package a task so another session, agent, or client can continue it, or when a new session must pick up earlier work, such as asking for current progress, the next step, or where a topic stands ("汇报一下当前进度", "下一步做什么", "接着上次做", "回忆一下 X 做到哪了") or to continue another agent's session. Writes a handoff document when asked or when wrapping up leaves work open; picks up from this project's recent Claude Code, Codex, and Kimi Code sessions, checked against the repository. Do not use for a plain end-of-task summary.

FileKindSize
SKILL.mdskill md10.1 KB
agents/openai.yamlconfig321 B
references/handoff-document.mdmarkdown929 B
references/status-card.mdmarkdown2.1 KB
scripts/session_digest.pyscript19.9 KB
scripts/test_session_digest.pyscript18.9 KB

references/status-card.md

View on GitHub
# Status Card

A status card answers, for a user who did not watch the work: what is mine to do now, what is done and how we know, and which decisions are mine. Keep it to about 40 lines so it is read in full. Write the headings and content in the user's language.

## Template

```markdown
# <Project> status (updated <date>)

## This stage
<Goal of the current stage in one line.>
Deadline: <date> (<days left>). Review or exit conditions: <conditions, or "none set">.

## Your turn
<First line: whether any agent is running now, and where.>
1. ★ <The one thing that matters most.>
2. <Other things for the user to do.>
3. Waiting for your decision: <each open decision, one line, with what it blocks>.
Agents next: <what agents will do, and when>.

## Done, and how we know
- <item>: <evidence level>

## Decisions
- <decision>. Why: <reason>. Who: <you | ⚠ agent, unconfirmed>. <date>

## Details
<Paths to the handoff, plan, and lists, for looking things up; not required reading.>
```

## Evidence Levels

Mark each done item with the strongest level that holds, and name the check when it is automated:

- **You used it**: the user tried it themselves.
- **Checks pass**: tests, builds, or scripted runs passed; name them.
- **Code read only**: someone read the code; nothing ran.
- **Not verified**.

An item the user has not tried says so, even when checks pass. Untried behavior the user must judge belongs under "Your turn".

## Decisions

- Record why when the reason would change how the user judges the decision later.
- Who decided is either the user or an agent. An agent's decision carries ⚠ until the user confirms it; a record that does not show the user's approval counts as the agent's.
- Keep rejected options in the handoff or plan; the card names only what was chosen.

## Staying Short

- Replace stale lines instead of adding below them. The card shows the present, not a history.
- When it grows past about 40 lines, move confirmed decisions that no longer affect the current stage into the document listed under Details. Do not compress the wording to make room.
- A fact that cannot be verified now is left out or marked unverified, never copied forward.
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexUnverifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall