Individual skill

autodl-autogpu

Use when work in any project needs an AutoDL instance powered on or off in GPU mode (有卡) or non-GPU mode (无卡), switched between them, checked (status, balance, billing, free GPUs), or kept from burning GPU time while idle or after the session ends (空转, 忘关机, 自动关机); also, where the project's GPU is an AutoDL instance, before experiments, training or data transfer run on it, e.g. in an automated research or experiment pipeline, under another experiment skill, or on a request like 按计划自动跑实验 (自动科研).

SKILL.md

Read pinned source

Source commit 5cd650f72fea3793357b15e3e8618cb10fc41d83

What was checked

Matches the observed repository HEAD: Yes
Scanner rules are current: No
Repository scan complete: No
Skill scan complete: No
Observed behavior: Not checked yet
Last checked: Nov 5, 2026

These are observations at the recorded commit, not a guarantee of safety. Review the source before running it.

Install this skill

Choose an agent in your Stack or copy a command below. Commands fetch the pinned commit and copy only this skill directory. Existing destinations are refused. Copying a command does not install anything.

An install command is unavailable: current rules, complete repository and skill scans, an eligible behavior level and a permitted license are required.

Repository context

Repository stars (not skill ratings): 95
Repository last updated: Oct 7, 2026
License: MIT

Observed agent integration: No specific integration evidence. Standard SKILL.md installation does not imply verified compatibility.

agent-skills Repository topic
autodl Repository topic
automation Repository topic
claude-code Repository topic
claude-skills Repository topic
deep-learning Repository topic
gpu Repository topic
data Derived category
research Derived category

Files and excerpt

File metadata and permitted, redacted excerpts come from the recorded scan. Source links are pinned to that commit.

Browse files

.gitattributesasset · 217 bytes
.gitignoreasset · 28 bytes
LICENSEasset · 1,064 bytes
README.cn.mddoc · 15,433 bytes
README.mddoc · 18,751 bytes
SKILL.mdcore · 38,972 bytes
SKILL.mdcore · 38,972 bytes
SKILL.mdcore · 38,972 bytes
SKILL.mdcore · 38,972 bytes
SKILL.mdcore · 38,972 bytes
assets/demo.cn.svgasset · 9,004 bytes
assets/demo.svgasset · 9,007 bytes
assets/hero.cn.svgasset · 9,557 bytes
assets/hero.svgasset · 9,551 bytes
assets/nogpu.cn.svgasset · 11,036 bytes
assets/nogpu.svgasset · 11,040 bytes
assets/parts.cn.svgasset · 15,119 bytes
assets/parts.svgasset · 15,127 bytes
assets/safety.cn.svgasset · 7,538 bytes
assets/safety.svgasset · 7,531 bytes
dev/analyze_samples.pyruns · 11,763 bytes
reference/clone-page.jsruns · 41,163 bytes
reference/clone-page.min.jsruns · 29,693 bytes
reference/clone.mddoc · 44,999 bytes
reference/console-clone.min.jsruns · 45,885 bytes
reference/console-more.mddoc · 12,676 bytes
reference/console.jsruns · 65,033 bytes
reference/console.mddoc · 42,580 bytes
reference/console.min.jsruns · 34,098 bytes
reference/guard.mddoc · 39,302 bytes
reference/ledger.mddoc · 36,544 bytes
reference/ssh.mddoc · 29,928 bytes
scripts/autodl_ctl.pyruns · 277,989 bytes
scripts/autodl_guard.shruns · 134,251 bytes
scripts/ctlruns · 1,437 bytes
tests/README.mddoc · 3,163 bytes
tests/conftest.pyruns · 1,728 bytes
tests/console/clone_fixtures.jsruns · 33,097 bytes
tests/console/clone_run.jsruns · 54,387 bytes
tests/console/fixtures.jsruns · 46,994 bytes
tests/console/fn_digest.pyruns · 1,222 bytes
tests/console/paste_copy.pyruns · 4,064 bytes
tests/console/run.htmlasset · 1,293 bytes
tests/console/run.jsruns · 105,453 bytes
tests/console/run_headless.pyruns · 7,845 bytes
tests/console/skeleton.jsruns · 23,549 bytes
tests/console/skeleton_test.jsruns · 36,560 bytes
tests/fixtures/autodl_guard_0.7.1.shruns · 57,950 bytes
tests/live/autostart_probe.conf.inasset · 269 bytes
tests/live/autostart_probe.profile.inasset · 701 bytes
tests/live/check_ini.pyruns · 764 bytes
tests/live/cpu_duty.pyruns · 995 bytes
tests/live/disk_write.shruns · 1,551 bytes
tests/live/export_logs.pyruns · 3,028 bytes
tests/live/fingerprint.shruns · 1,247 bytes
tests/live/gpu_duty.pyruns · 1,219 bytes
tests/live/gpu_load.shruns · 295 bytes
tests/live/jobs_done.pyruns · 592 bytes
tests/live/probe_state.pyruns · 2,076 bytes
tests/live/queries.shruns · 3,343 bytes
tests/live/sshread.shruns · 2,723 bytes
tests/live/ticket_probe.pyruns · 6,309 bytes
tests/local_tools.pyruns · 2,017 bytes
tests/store_posix_check.pyruns · 6,203 bytes
tests/stubs/fake-screenruns · 1,331 bytes
tests/stubs/nvidia-smiruns · 1,770 bytes
tests/stubs/screenruns · 539 bytes
tests/stubs/shutdownruns · 703 bytes
tests/stubs/syncruns · 954 bytes
tests/stubs/tmuxruns · 268 bytes
tests/test_analyze_samples.pyruns · 9,381 bytes
tests/test_clone.pyruns · 65,275 bytes
tests/test_clone_remote.pyruns · 29,051 bytes
tests/test_console_static.pyruns · 25,195 bytes
tests/test_ctl.pyruns · 111,125 bytes
tests/test_docs_static.pyruns · 40,467 bytes
tests/test_guard.shruns · 236,995 bytes
tests/test_live_scripts.pyruns · 29,136 bytes
tests/test_local_tools.pyruns · 5,644 bytes
tests/test_store.pyruns · 96,931 bytes
tests/test_ticket.pyruns · 28,786 bytes
tests/test_transfer_wsl.pyruns · 20,442 bytes

Read excerpt


# AutoDL 开关机

本 skill 让当前对话里的 AI 自己完成 AutoDL 普通容器实例的有卡开机、无卡开机、关机与模式切换，不用人去点。它只提供方法。什么时候开、开哪种模式、什么时候关，由你按任务和用户的习惯决定，文中的例子只作参考。三样东西配合着用。

- **你是主控。** 开机在控制台网页上做，照 `reference/console.md`（下称手册；第 13、15、16、17 节在 `reference/console-more.md`）一步步来；其余经本机助手 `ctl` 走 SSH
- **守护程序兜底。** 实例上的守护按实际占用判断空闲，空闲满用户定的时长就关机。你的会话断了或忘了关，它照样关
- **控制台定时关机是可选的硬上限。** 用户选了才设，到点由平台直接关机，会切断任务。不由用户选的只有开机流程第 2 步的临时定时

对用户少打扰，只汇报结果（见"汇报"）。不要另写守护脚本或 `sleep; shutdown` 之类的兜底。

## 平台事实
- 普通容器实例没有开机 API，开机只能在控制台网页上点。不用任何令牌或非公开接口；只有用户开启了自动克隆时，页面脚本才从页面自己的数据里读新实例的 SSH 主机与端口，不读密码
- 在实例里执行 `/usr/bin/shutdown` 就关机并停止计费，由守护程序与 ctl 调用
- 按量计费从开机到关机按秒算，与是否用 GPU 无关。关机后 GPU 不预留，再开有卡时主机可能没有空闲 GPU。数据保留到实例被释放