Borrowing it
Nothing to install: this file belongs to myths-labs/muse. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/myths-labs/muse/main/.claude/commands/handoff.mdgit clone --depth 1 https://github.com/myths-labs/museWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/myths-labs/muse/handoff)<a href="https://agentmods.dev/commands/myths-labs/muse/handoff"><img src="https://agentmods.dev/badge/commands/myths-labs/muse/handoff/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/myths-labs/muse/handoff"><img src="https://agentmods.dev/badge/commands/myths-labs/muse/handoff.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00052 | $0.01872 |
| Opus 5.5 | $0.00021 | $0.00749 |
| Sonnet 5.5 | $0.00010 | $0.00374 |
| Haiku 4.5 | $0.00005 | $0.00187 |
Grade A, and why
handoff scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 52 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/handoff — 账号接力(多线编排,可选)
会话不能跨账号搬;检查点、看板、记忆目录和本机定时任务都在本机,新账号开新会话读文件就能接上。本流程把「存」和「接」固定下来。
前提:项目有 .muse/orchestration.json(模板 .agent/skills/core/muse-orchestration/templates/orchestration.example.json),登记编排者(orchestrator,即「总管」)和用户自定义的各条线:线 id、检查点文件、接手必读。没有这个文件时按单线处理:存用 /save,接用 /resume。完整规则与依据见 muse-orchestration/references/ACCOUNT_RELAY.md。
用法
/handoff save:在旧账号的会话里执行,写本线检查点和接力快照。/handoff <线 id>:在新账号新开的会话里执行,例如/handoff orchestrator、/handoff build。/handoff status:只读,列出各线最近一条快照和距今多久。
什么时候存
- 谁都可以读额度:客户端的用量读数是账号级的,任何会话都能读,不必等编排者。
- 滚动检查点:每派出或收回一批子代理、每关一个包,就更新本线检查点的「当前态」(在跑哪些子代理与分支、下一步、在等谁)。撞墙时最多丢一批。
- 存档线只看周额度:用最近两次读数算「预计用完分钟数」= 剩余百分点 ÷ 每分钟消耗。预计 20 分钟内用完,或周额度 ≥95%,任一先到 → 立刻
/handoff save并告诉编排者;编排者收到第一条就向全部线发「全局保存」。存完照常干到撞墙,撞墙才切号。 - 短窗(如 5 小时窗)快满时不发全局保存、不必 save:它会自己重置、不换账号,靠第 2 条的滚动检查点。每条线始终留一条「短窗重置时刻 +1 分钟」的自续定时,提示「照进度文件最后一行续」;撤停派令或提前开工时也不删它。
- 一次续派 ≥10 个子代理之前先写一次检查点。
- 新账号先看档位:额度明显更小的档位全线开会很快撞墙,先告诉用户,等他定再开其它线。
- 不按上下文百分比换会话:各线让客户端自动压缩后继续,上下文到 80% 时增量更新一次检查点,不停手、不开新会话。只有三种情况才
/handoff save并换新会话:①切账号(全线一起存);②会话明显糊涂(反复读错状态、压缩后关键状态互相矛盾);③撞墙前预计要空闲 60 分钟以上、上下文又超过 30 万 token——先存,重置后由新会话从检查点接手,不叫醒旧会话。同一线里要同时做的活交给本线的子代理,不多开会话。
/handoff save 做什么
date取时间,写进检查点标题。- 先写记忆再写检查点:只把本会话新出现的用户决定(引原话 + 时间)写成记忆(一条事实一个文件,并更新索引);评估结论、对比数据写进检查点即可。写前用
recall.py <词> <其它语言同义词>查重,已有就改原文件。这里不跑memory_audit.py --sync(见第 5 条)。 - 写检查点:在做什么、做到哪一步、下一步、在等谁、未提交或未落盘的东西放在哪、在跑的子代理及其进度文件。未提交的改动提交到本线自己的分支,不推主干。
- 在接力手册(配置
relay_runbook)的「交接快照」填本线一行;编排者另填主干提交号、线上版本、等用户定的事。 - 存档只写三样:本线检查点、接力快照、新的用户决定(记忆)。Obsidian 库(
daily/、decisions/)、vault_coverage_check.py与memory_audit.py --sync不在每次 save 里做,由编排者每天晚报统一跑一次(见muse-orchestration/references/ORCHESTRATOR.md「Daily maintenance」与VAULT.md)。 - 编排者退役前删掉自己的会话内定时(应用重启会恢复旧会话的定时,到点会把退役会话叫醒)。
- 最后一句告诉用户:「<线 id> 快照已写,可以切」。
/handoff <线 id> 做什么
date校时;按配置读该线的「接手必读」,不读完不说话。- 读额度:记下档位、短窗与周百分比。若本线最后检查点比旧会话撞墙时间早 10 分钟以上,先跑
wall_recover.py <旧会话 id> --since <最后检查点时间>,把没跑完的子代理与各自最后进度补写进检查点,再续派。 - 解开上级会话:本会话若是别的会话开出来的(有上级且未解开),先解开成顶层会话;否则旧会话一归档会把本会话一起归档。
- 列出其它会话,把自己的新地址发给各方(第一行写「<线 id> 已在新账号接上,之后发这里」);同名会话用唯一 ref 区分。每个收件人只发一条。客户端对跨会话消息设有上限时,到上限就把没发的写成待发草稿,等用户下一句再发,不换通道绕过(见
ORCHESTRATOR.md「Cross-session message cap」)。 - 核对快照里「在等谁」是否已有结果(问对方或只读查),不照搬旧状态。
- 向编排者发第一条报告(编排者向用户):最要紧三件各到哪一步、要定的事不超过三条、每条带推荐。
- 编排者另做:按
muse-orchestration/references/ORCHESTRATOR_CRON.md重设定时;核旧编排者留下的停派令,条件已过的当场撤;旧会话改名「已退役(勿发消息)」或归档,并请它删自己的定时。 - 按快照的「下一步」继续。权限边界不变:部署、生产数据变更、付款、对外发送、删除数据仍需用户在本会话明确同意。
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 5d ago Changed 2633ee7b237b
- 12d ago First seen · 52 lines · 52 tokens per session scan A ce62dcd565e9
handoff is a command published in the GitHub repository myths-labs/muse (38 stars, last pushed 6d ago), licensed MIT. It adds 52 tokens to every session and 1,872 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-25.
handoff compared with commands from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| handoffmyths-labs/muse | Command | 38 | 6d ago | MIT | A | 52 | Claude Code |
| handoffrosehgal/handoff | Command | 2 | 4mo ago | MIT | A | 10 | Claude Code |
| dejavshulcz/deja-vu | Command | 1.4k | yesterday | MIT | A | 17 | Claude Code |
| adhoc-flowgriddynamics/rosetta | Command | 370 | today | Apache-2.0 | A | 22 | Claude Code |
| prepsuboss87/FDEOps | Command | 857 | 17d ago | MIT | A | 10 | Claude Code |
| orchestrateegorfedorov/claude-context-optimizer | Command | 130 | 10d ago | MIT | A | 0 | Claude Code |
| budgetcukas/remembrall | Command | 6 | 3d ago | MIT | A | 13 | Claude Code |