Release Ota Promote logo

Release Ota Promote

OrganizationPopular
Prismer-AI
release-ota-promote

Publish signed Electron UI/daemon bundles to a local OTA feed with atomic manifest switch and digest readback, while dry-running the K8s drain_respawn rollout. Prod is refused before writes; feed publication is not client apply.

Overview

PublisherPrismer-AI
RepositoryPrismerCloud
Skill namerelease-ota-promote
Stars
1.6K
Forks
17
Bundled files
1
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 1 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by Prismer-AI on GitHub. Read the source before you install it.

Installation

Install the Release Ota Promote AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/Prismer-AI/PrismerCloud/tree/main/sdk/apc/skills/release-ota-promote
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/Prismer-AI/PrismerCloud.git /tmp/PrismerCloud
mkdir -p .claude/skills
cp -r /tmp/PrismerCloud/sdk/apc/skills/release-ota-promote .claude/skills/release-ota-promote
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Release Ota Promote in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Release Ota Promote on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Release Ota Promote is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

release-ota-promote

把一份 runtime OTA 沿双通道下发(apc/01 §2 ota-promote,收口步 7/8/11)。这是线 3 admin fleet 下发,交付模式治 doc01 #8("promote 用了线 1 kill1 而非线 3 drain_respawn"):

  • K8s daemon 通道:交付 drain_respawn(排空在飞 task → 优雅 exit(0) → supervisor respawn → boot-OTA 换版,用户无感/无中断)。正规落点 = admin fleet rollout endpoint POST /api/admin/runtime-releases/rollout(src/im/api/admin.ts)→ 给每个目标 daemon 下发 runtime.update.apply 帧 → daemon arm drain_respawn。本 verb 不再 spawn 线 1 kill 1 的 runtime-ota.ts(那会打断在飞 task),而是构造这条 drain_respawn rollout 请求。本机替身只 dry-run 出请求、不真 POST(真 POST 需 approvalId + release row + 在线 daemon = M5)。
  • 桌面 feed 通道:本机替身未接线——诚实标 ⬜(铁律 4)。缺①桌面本地 feed 基座(apc/05 §3,P0-4 未落)②本机禁写真 Nacos manifest(= M5)。

什么时候用:一个新 runtime 版本发布后,要让 fleet 里的 daemon 采用它。

承重纪律(发版 skill 的命根子,先记死)

  • prod OTA 必拒(不变量 2)。--env prod 先于任何 OTA 动作硬拒,不做任何下发。别试 prod:真 prod promote = M5。
  • 交付模式 = drain_respawn,绝非 kill1。线 3 fleet 下发必须排空在飞后再 respawn(kill 1 会打断用户在飞 task,那是线 1,仅 dev/紧急)。k8s.directive 必须是 drain_respawn。
  • 不宣称 drain_respawn 已执行。本机替身只 dry-run 出 rollout 请求(k8s.posted:false),不真 POST(= M5:approvalId + release row + 在线 daemon)。照实报"已构造 drain_respawn 请求,未真下发",绝不说"已 drain_respawn 下发"。
  • 桌面 feed 走真文件副作用。传 --desktop-feed-dir 与 UI/daemon builder metadata;verb 先校验 size + SHA-512,复制版本化 bundle,最后原子切 manifest.json,再从磁盘回读摘要。只有 desktop.status:'applied' + readback.verified:true 才能宣称 feed 已发布。
  • feed 已发布 ≠ Electron 已应用。客户端 check → stage → apply 需要另取运行态证据,不得从文件写入外推。

Anti-pattern(直接判红)

  • ❌ --env prod 想 promote prod。命中即 blocked,别绕。
  • ❌ 把交付模式报成 kill1 / 线 1 kubectl exec kill 1。本 verb 是线 3,交付 drain_respawn;报 kill1 = 把在飞打断当成了正规下发。
  • ❌ 把 k8s.posted:false 的 dry-run 说成"已 drain_respawn 下发/已 POST",或把本地 feed 已发布说成"Electron 已更新"。

工具契约(签名以此为准)

apc 不在 PATH——在仓库根用 npx tsx sdk/apc/bin/apc.ts <sub>。

命令作用退出码
npx tsx sdk/apc/bin/apc.ts release ota-promote [--env dev|test|prod] [--json]K8s 通道构造 drain_respawn rollout 请求(落点 admin fleet rollout endpoint,本机替身 dry-run 不 POST)+ 桌面通道 ⬜;prod 硬拒0 green(真 rollout 完成,仅 M5)· 1 blocked(本机 dry-run 边界 / 桌面未接线 / prod 拒)

--json 字段(名以此为准):

  • 顶层:{ verb:'ota-promote', decision, env, deliveryMode, k8s, desktop, blockers:[], notes:[] }
  • k8s:{ channel:'admin-fleet-rollout', endpoint, directive:'drain_respawn', frame:'runtime.update.apply', request:{channel,version,target}, posted:false, boundary }(directive = 交付模式,必须 drain_respawn;posted:false = 本机替身未真 POST)
  • desktop:成功时 { status:'applied', deliveryMode:'local-file-feed', plan:{manifestPath,...}, readback:{verified:true,...}, boundary };校验/写入/回读失败时 {status:'blocked', blockers:[...]}。
  • deliveryMode:'drain_respawn(线 3 admin fleet rollout · 本机替身 dry-run)'——不是 kill1

Workflow

1. K8s 通道(构造 drain_respawn rollout 请求)

bash
npx tsx sdk/apc/bin/apc.ts release ota-promote --env dev --json > /tmp/ota.json; O=$?
echo "ota-promote exit=$O"
  • 读 k8s.directive——必须是 drain_respawn(线 3,排空在飞后 respawn),不是 kill1。
  • 读 k8s.posted——false = 本机替身只 dry-run 出请求、未真 POST(真 POST = M5)。
  • 本机替身 → decision:blocked(exit 1)+ blocker 说明"真 admin rollout POST = M5"。如实报,这是本机替身边界(release-common.ts 本机替身边界),不冒充已下发。

2. 桌面通道(诚实 ⬜)

使用隔离目录与现有已签名 builder metadata 发布并回读:

bash
npx tsx sdk/apc/bin/apc.ts release ota-promote --env dev \
  --desktop-feed-dir .e2e-tmp/apc-ota-feed \
  --desktop-ui-meta apps/desktop/dist-ui/ui.manifest.json \
  --desktop-daemon-meta apps/desktop/dist-ui/daemon.manifest.json --json

核验 desktop.status=applied 、desktop.readback.verified=true,并独立检查 desktop.plan.manifestPath 及版本化 bundle 存在。

3. prod 一律拒(验证人闸)

bash
npx tsx sdk/apc/bin/apc.ts release ota-promote --env prod --json; echo "prod exit=$?"
# 预期:decision=blocked, exit 1, k8s.skipped=true, blockers 含 "prod 人闸"

Failure / 边界

  • 本机替身 → K8s 通道 dry-run 出 drain_respawn 请求但不真 POST → blocked。这是预期的真行为(本机替身边界),不是 skill 的红。
  • 桌面通道 = ⬜(feed 基座未落 + 本机禁写真 Nacos)。
  • 真 drain_respawn fleet 下发(真 POST admin rollout endpoint + approvalId + release row + 在线 daemon)= M5;本 verb 只把交付模式纠正为 drain_respawn 并 dry-run 出请求。

输出契约(机器判据按这个复算,别自由发挥格式)

本 skill 的判据不是「报告里出现了 drain_respawn 这个词」——旧判据正是一条 match:"drain[\s_-]?respawn",一份宣称 posted:true、decision:green、"fleet 已换版"的纯编造报告在它下面满分。新判据重解析你贴的两份 --json 产物,把 k8s.channel/directive/frame/posted 逐字段钉死、从 decision 反推退出码、并把 rollout body 的 version 与磁盘上真 /VERSION 比对(json-claim)。

报告必须带这组行 + 两份原样 JSON(--env dev 在前,--env prod 在后):

DIRECTIVE: drain_respawn
POSTED: false
DESKTOP: applied
DESKTOP-VERIFIED: true
RUN-EXIT: 1
PROD-DECISION: blocked
PROD-K8S-SKIPPED: true
PROD-RUN-EXIT: 1

判据会判红的情况(任一):

  • 没有可解析的 fenced JSON;
  • k8s.directive 不是 drain_respawn(写成 kill1 直接红——线 1 会打断在飞 task,把它当交付模式正是本 verb 存在的理由)、或 k8s.channel != admin-fleet-rollout、k8s.frame != runtime.update.apply;
  • k8s.posted != false,或 posted:false 却 decision:green——什么都没 POST,就不许把任何东西报成已下发;
  • desktop.status != applied 或 desktop.readback.verified != true,或把 feed 发布冒充为 Electron 客户端已 apply;
  • DIRECTIVE / POSTED / DESKTOP 与产物字段不一致,或 RUN-EXIT 不是 decision 推出来的;
  • rollout body 的 k8s.request.version 不等于磁盘上 /VERSION;
  • prod 产物不是 blocked + k8s.skipped:true(连 rollout 请求都没构造)+ blockers 非空。

诚实边界:桌面通道有真文件副作用与摘要回读;K8s 通道仍只构造请求、不 POST。因此顶层仍应 blocked,直到真 admin rollout 产生送达证据。

产出(副作用 oracle,报告里必须给)

  1. decision + 真退出码:green/blocked ⇄ 0/1。
  2. 交付模式证据:k8s.directive:'drain_respawn'(线 3,非 kill1)+ k8s.frame:'runtime.update.apply' + k8s.channel:'admin-fleet-rollout'。
  3. 本机替身边界:k8s.posted:false(未真 POST)+ blocker 说明真 POST = M5。
  4. 桌面 feed 真回读:desktop.status:'applied' + readback.verified:true + manifest/bundle 路径。
  5. prod 拒证据:--env prod → blocked + exit 1 + "prod 人闸" + K8s 通道 skipped。

不许:把交付报成 kill1;把 dry-run(posted:false)说成已 drain_respawn 下发;把 feed 发布说成 Electron 已 apply;试 prod。

Checklist

  • 跑了 apc release ota-promote --env dev --json,k8s.directive = drain_respawn(非 kill1)
  • k8s.posted:false 时如实报"已构造 drain_respawn 请求、未真 POST(= M5)",没冒充已下发
  • 桌面 feed 照实报 applied + readback.verified + manifestPath,没冒充 Electron 已 apply
  • delivery 明确是 drain_respawn(线 3 admin fleet rollout),没报成 kill1
  • 演示 --env prod 被拒(blocked/exit 1),没试 prod

审批门(doc03 §1)

本 skill 收口 doc03 §1 两个审批点:「OTA promote/下发」(ota-promote ga/rollout 前)与 「桌面通道发布」(桌面 bundle/manifest 指针更新前,收口步 11)。dry-run 停在下发边界,真下发前必须建人审批并停手:

  1. 出示双通道证据(审批人看到的):K8s = dry-run request + target/version;桌面 = 已原子发布的 manifest + bundle 签名/摘要回读。

  2. 建真 cloud 审批(真路径 = human-approval built-in skill 的 CLI)——把上面证据装进请求:

    bash
    cloud approval request-human --task-id "$PRISMER_TASK_ID" \
      --action "OTA rollout runtime <version>(drain_respawn,test fleet)" \
      --context "target daemon: <k8s.request.target>;版本分布见 dry-run 回执;桌面通道 manifest 指针 diff / 签名摘要(未接线 ⬜)" \
      --risk "线 3 fleet 下发;drain_respawn 排空在飞后 respawn(非 kill1 打断在飞 task)"

    返回 approvalId;建完即停手本轮(human-approval 铁律:平台在人裁决后重派)。

  3. block:桌面 feed 可已验证,但 k8s.posted:false 仍使顶层 blocked(exit 1)。

  4. post-approval = M5 real-remote:真 POST /api/admin/runtime-releases/rollout(消费 approvalId + release row + 在线 daemon)属 M5 real-remote,本机替身不执行、不冒充已下发。

PKF 直写(研发回环 · doc07 §B2 写入点①)

产出时机:ota-promote 出 decision 后,把这次 OTA 下发的结论直写成一张 PKF 记忆页(pageType=decision,ReleaseCard 语义 apc/09 §1)——同一页既是回审批人侧的富报告,也是可召回的发版记录。经验教训另走 memory skill 的 CONSTRUCT→PLACE→WRITE,两条边并存(doc07 §B1)。

语法照 pkf-writing skill——本 skill 不再内嵌语法骨架(frontmatter / typed link / 数据块写法都在那边)。写入面不新造:code agent 走 prismer memory write(SS-14 §4.3 appendix;hermes 侧是 native memory_write)。

声明(doc10 §2.5 格式,报告末尾一行):

PKF: prismer://workspace/<ws>/memory/<path>

写入面不可达时如实声明 PKF: none — 写入面不可达(<哪一条>),绝不允许为了满足规则而假装写了。

回读:声明后必须回读该页(pkf_read / memory read),确认真实存在、正文非空且与结论一致(k8s.posted=false 就照实标未下发,桌面仅标 feed 发布与回读,绝不写成"已 drain_respawn 下发"或"Electron 已 apply");typed link 的 prismer:// 目标必须真实存在(validator 规则 2c/4 挡 invalid-prismer-host),无对应页就删该 link 行、宁缺勿造伪目标;写入自动挂 INDEX 反孤儿锚(SS-14 §4.2),优先 edit 既有页而非 dump 新叶(PLACE 治理照走)。

调用回执(APC 平台方运营纪律 · 自动注入,勿手改)

执行本 skill 的第一步,先落一条调用回执:

bash
cloud skill ack release-ota-promote --task "$PRISMER_TASK_ID"

按退出码分流(这条命令的退出码是承重信息,禁止用 || 兜底、; true、 set +e 或重定向把它抹掉):

exit含义你要做的
0回执已落库(im_task_logs.action='skill_ack')继续执行本 skill
3无 task 上下文——本次运行没有 task,产不出回执继续执行本 skill;但本次运行没有回执,任何报告里都不得声称已 ack
4你不是该 task 的 assignee,服务端拒绝停下并上报:回执只能由执行该 task 的 agent 产生
1其它失败(网络 / 服务端)重试一次;仍失败则继续执行,并在结果里显式标注「回执缺失」

回执只证明本 skill 被调度,不证明执行正确——效果证明由本 skill 自己的 acceptanceCriteria 副作用断言承担(apc/04 §2 层 1 诚实标注)。

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Release Ota Promote AI skill do?

Publish signed Electron UI/daemon bundles to a local OTA feed with atomic manifest switch and digest readback, while dry-running the K8s drain_respawn rollout. Prod is refused before writes; feed publication is not client apply.

Why use Release Ota Promote on TypingMind?

Because you install it once and use it with any model. Release Ota Promote is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Release Ota Promote in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/Prismer-AI/PrismerCloud/tree/main/sdk/apc/skills/release-ota-promote. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Release Ota Promote?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Release Ota Promote?

As many as you like. As long as a model supports skills, you can use Release Ota Promote with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Release Ota Promote AI skill free?

Yes. It is published on GitHub by Prismer-AI under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇