
alibabacloud-cms-manage
Entry skill for the aliyun CLI distribution of CloudMonitor (CMS). Use when the user mentions aliyun cms2, CloudMonitor, CMS commands, or any CMS module operation such as Integration Policy/Center, APM, RUM, Prometheus Service, Recording rule, alert rule, alert template, alert history, event hub, SLS event, PromQL, cloud resource, service observability, monitoring onboarding, metric query, etc.
- Installs
- 245
- Last 24h
- —
- Trending rank
- —
- Security
- Not audited
Install
Works with every agent the skills CLI supports.
npx skills add https://github.com/aliyun/alibabacloud-aiops-skills --skill alibabacloud-cms-manage- 01
Install
Run the command in your project. The CLI asks which agents to install for.
- 02
Describe the task
Ask your agent for something this skill covers. It loads alibabacloud-cms-manage automatically when relevant.
- 03
Review the result
The agent follows the SKILL.md steps, scripts and references bundled with the skill.
SKILL.md
The instructions your agent reads when the skill is triggered.
CMS CLI — aliyun cms2
Prerequisite Check
Once per session: perform this prerequisite check only on the first invocation of this skill in a conversation. If all checks already passed earlier in the same session, skip directly to the relevant module.
-
Check
aliyunexists —which aliyun(macOS/Linux) orwhere aliyun(Windows).- Not found → ask the user to install the aliyun CLI first: https://help.aliyun.com/document_detail/121541.html. Stop and wait.
-
Check CLI version — run
aliyun version. Minimum required: 3.3.15 (seecompatibilityin frontmatter).⚠️ Compare version segments as integers (semver): 3.3.4 < 3.3.15 because 4 < 15. Shell verification:
printf '%s\n' "3.3.15" "$(aliyun version)" | sort -V | head -1If the output equals the current version, the requirement is NOT met.- Version OK → go to step 3.
- Version too old or unrecognized →
- Run
aliyun upgrade --helpto test whether theupgradesubcommand exists.- Available → run
aliyun upgrade -yto update to the latest version automatically, then re-checkaliyun version.
- Available → run
- If
upgradenot available → ask the user to upgrade manually following the official guide: https://help.aliyun.com/zh/cli/update-cli. Wait for the user to confirm the upgrade is complete, then re-checkaliyun version. - If upgrade succeeded → go to step 3.
- If upgrade failed → ask the user to upgrade manually: https://help.aliyun.com/zh/cli/update-cli. Stop and exit.
- Run
-
Check
cms2plugin — runaliyun cms2 --help.- Help output OK → continue to Credentials.
unknown command/ missing → stop immediately, output the error report below (append CLI version, OS, and error message), and make no further CLI calls.
Credentials
aliyun cms2 reuses the aliyun CLI credential system (aliyun configure).
Use --profile <name> to switch profiles.
Observability
Upon loading this skill, generate a random session ID (32-char lowercase hex string) once per session...
For every aliyun cms2 command, append: --user-agent "AlibabaCloud-Agent-Skills/alibabacloud-cms-manage/{session-id}"
Local utility commands (e.g. version, configure, plugin, --help) do not support this flag and are excluded.
Global Conventions
Hard constraint: fallback to aliyun cms, other API versions, or any workaround is strictly prohibited.
Run
aliyun cms2 <command> [subcommand] --helpbefore first use of a subcommand in a session to get the full flag list and examples. Once the help for the same subcommand has been read in the current session and the command shape has not changed, reuse that knowledge instead of repeating the help call. A named subcommand's--help/--show-schema/--show-example-bodyis the authority for that command's flags, body envelope, response fields, and environment behaviour — do not search other skill files for a copy of those fields.
- Prefer
-o text(default) to reduce token consumption for list/get; use-o jsononly when the output is parsed field by field rather than read. - Addon-release
values: expand each field'sfieldPathinto nested JSON as specified in Addon Values Defaults. Do not write a dottedfieldPathas one literal key. How to pick--env-type, where child fields sit on create vs update, and when a subset is (or is not) a valid body, are in that section. - Region parameter required for mutating and detail commands: unless otherwise specified by a module-specific rule, all
create,update,delete, and single-resourceget/detail commands MUST include the--regionparameter so the request is routed to the correct backend OpenAPI endpoint. Omitting--regionon these commands may cause routing failures or operate against an unintended region, and the error can name something else entirely —integration policy createreturnsstatus 400: The workspace can not be createdeven when the workspace exists and the body is valid. This does not apply tolist/query commands that intentionally span multiple regions (e.g.entity query --source CloudResourceall-region queries). Requiring the flag is not permission to choose its value — resolve it per Region Confirmation Gate. - One choice, one question: every enumerated choice (region, workspace, policy, addon, scope mode, tag match mode, and any other mutually exclusive parameter) is asked as one question carrying all of its options. Never split them across questions or into a
(续)...continuation, and never drop the ones that do not fit — a split turns one choice into two answers that can conflict or be left incomplete, and a truncated list hides valid choices entirely. When the structured input form cannot render every option, ask as plain text and spell out every option with its explanation in the question body. A mutually exclusive choice stays single-select; only a genuinely multi-valued one (several regions) is multi-select. A prompt may carry several distinct choices, each its own question. Put any recommended value directly in the option label (for example(Recommended)); recommendations are advisory only and do not permit cloud-side writes unless another rule permits defaulting or the user confirms. - Human confirmation required for writes and high-impact creates: before any command that creates or changes cloud-side state (
create,update,delete,patch,start,stop, etc.), show a concise confirmation summary and the exact command, ask whether the user confirms execution, and wait for a clear affirmative answer. The summary must include the operation, target resource identifiers, expected impact, and notable risks or irreversible effects when applicable. Do not require an exact phrase or long confirmation text; a clear affirmative answer such as "yes", "confirm", "proceed", or "确认" is sufficient approval. Skip confirmation only for dry-run, preview-only, or read-equivalent creates with no cloud-side impact; if uncertain, require confirmation. - Uncertain parameters must be explicitly answered by the user: for any parameter whose value is not explicitly provided or cannot be reliably determined (e.g.
region/regionId, workspace, policy,resourceGroup, tag, resource scope,addon/addonName, resource type, cloud product/service name, onboarding configuration options, etc.), ask the user for a clear answer before proceeding. Never fabricate, guess, infer from defaults/history, or arbitrarily choose one value. A module-specific reference may define narrow exceptions — for the integration module they are listed in references/integration-common.md. Changing an existing addon release's settings is an unrehearsable write: follow Addon Release Config Update. - A discovery query is not an answer: a read command (
workspace list,entity query,policy list,sts get-caller-identity, etc.) only builds the candidate set you present, and its output never becomes the user's answer — not when it returns a single candidate, not when a naming convention makes the value derivable, and not when an earlier turn of the session used one. "Reliably determined" means the user's own words or the resource they named pinned the value, not that the query happened to be unambiguous. - Name-to-ID lookup must match exactly: when looking up a region ID, workspace ID, integration policy ID, resource group ID, resource type value, or cloud product/service name/code by name, if no exact match is found in the query results, do not silently pick an arbitrary value as a substitute. Instead, report the mismatch to the user and ask them to confirm or provide the correct value.
Region Confirmation Gate (Hard Requirement)
Applies to every module and to every command that takes --region or a region/regionId body field. Settle the region before the workspace, since workspace verification is region-scoped.
List/query commands omit --region only when a module rule says the call is all-region. Omitting the flag is not all-region: the CLI supplies the profile default. Mutating and detail commands must carry it per Region parameter required.
There are exactly three legitimate sources for the value:
- The user stated it in the current request — not an earlier unrelated turn of the same session.
- The user picked it from a candidate list you presented, and you waited for that answer before running the next command.
- The host injected it as runtime context (the console session's current region). Say which region you are operating on before using it.
Anything else is a violation, including: a CLI profile / aliyun configure / ALIYUN_REGION default, or an omitted required --region; the cn-hangzhou of docs and examples, or a region only stated in narration; splitting default-cms-{userId}-{regionId} (or any workspace name); passing controlRegionId.
A module may bind --region to the regionId of a resource the user already confirmed (named cluster, workspace that already passed this gate, instance). That is still Source 1 or 2. Module references do not loosen this.
If Source 1 and 2 are absent, present the region as a choice per One choice, one question and stop. Gather candidates with aliyun cms2 meta regions and offer every regionId (never showName or controlRegionId); only consuming that output is wrong. Source 3, when present, is a recommended option only — it routes the call and does not decide scope. Onboarding region scope is collected by Resource Scope Selection Gate; do not ask a second region question here.
controlRegionId is the control-plane region and may differ for dedicated / exclusive locations — never pass it as --region.
Workspace Confirmation Gate (Hard Requirement)
Applies to every module and to every command that takes --workspace or a workspace body field. There are exactly three legitimate sources for the value:
- The user stated it in the current conversation. Verify it exists in the target region by exact
workspaceNamematch; no exact match → report and ask, never substitute a near match. - The user picked it from a candidate list you presented, and you waited for that answer before running the next command.
- The environment supplied it as runtime context (e.g. the console session the skill runs inside) and its region matches the target region. Say which workspace you are operating on before using it.
Anything else is a violation: adopting the single row aliyun cms2 workspace list returned, picking the default-cms-{userId}-{regionId} entry because it looks like the account default, assembling that name from aliyun sts get-caller-identity plus a region, or reusing a workspace from an earlier unrelated request in the same session.
So aliyun cms2 workspace list remains the right way to gather candidates — only consuming its output instead of presenting it is wrong. Mark the best candidate (Recommended) inside the option per One choice, one question, then stop and wait.
Module references do not loosen this. Where one gives default-cms-{userId}-{regionId} as the workspace's "default format" or builds it from an account ID and a region, that describes how the account's default workspace is named — it is the candidate to recommend, never permission to skip the question.
Pagination & Query Failure Handling
- Paginate to completion: for every
listcommand that supports--next-token, keep querying until nonextTokenremains or accumulated count ≥totalCount. Do not trust the first page as complete when pagination metadata indicates more data. - Page size on paginated
list: when the command accepts--max-results, pass100and then paginate to completion. This document wins over--help: do not rely on the omitted-flag default.policy list --helpsays 30 (max 100) but omitting it can returnmaxResults: 100;prometheus instance listomitted default is 30. 20 pages × 30 can truncate a large list. - Accumulated count ≥ totalCount: stop even if
nextTokenis non-empty ortruncated=true;totalCountis the stronger signal. - Empty page with satisfied totalCount: stop even if
nextTokenis present. - Token loop protection: track seen tokens; stop on repeat and report as partial.
- Page limit: default 20 pages; report partial if reached.
- Truncated results: do not conclude absence from partial results. Use filters (
--search,--query,--policy-name, etc.) or paginate fully. - Transient query failure: retry once on a transient server-side error (
DEADLINE_EXCEEDED, timeout); if it still fails, mark asUnknown/QueryFailed— do not treat as healthy or unhealthy.
Error Handling
Error codes and actions are listed in aliyun cms2 --help. Additional tips:
InvalidJSONusually means malformed--body; validate withjq . <<<'<value>'before passing to the CLI.--body and stdin are mutually exclusive; specify only one— means both--body(or--file) and stdin data were provided. Fix: keep only one input source. In agent/CI environments where stdin may be a pipe, append< /dev/nullto the command to ensure stdin is empty.
Output Language and Terminology
- Write user-facing explanations, analysis, recommendations, summaries, and conclusions in the user's language; default to Simplified Chinese when that language is unclear or mixed. Follow a mid-conversation switch from that point on, and let an explicit instruction about output language override both.
- Question prompts, option labels, table headers, and reports are user-facing text too — phrase them in the answer language even where a reference file spells them out in one language.
- CLI command names, flags, API paths, JSON field names, enum values, resource IDs, metric names, and log/error messages MUST remain verbatim English/code, whatever the answer language.
- Answering in English: use the Glossary's English column as the canonical vocabulary rather than inventing synonyms.
- Answering in Chinese: use the Glossary's Chinese terms in all prose, never leaving a mapped term in English. Write
中文(English)on first mention only when it disambiguates, then the Chinese term alone. Before sending, scan for mapped English terms and replace them, except inside code, commands, JSON fields, IDs, or quoted CLI output.
Examples (Chinese answers):
- Good:
接入配置(AddonRelease,CLI 命令为 addon-release) - Good:
查询接入配置状态:aliyun cms2 integration addon-release list ... - Bad:
all releases are Ready - Better:
所有接入配置均 Ready
Glossary
| English | 中文 |
|---|---|
| Cloud Monitor / CMS | 云监控 |
| Workspace | 工作空间 |
| Application Monitoring / APM | 应用监控 |
| RUM | 用户体验监控 |
| Synthetic Monitoring / Synthetic | 云拨测 |
| CloudResource | 云资源 |
| EntityStore | 实体仓库 |
| Entity | 实体 |
| Integration Policy / policy | 接入策略 |
| Addon / addon | 组件 |
| Addon Catalog | 组件目录 |
| AddonRelease / addon release / release | 接入配置 |
| Collector | 采集器 |
| Prometheus View | Prometheus 聚合视图 |
| AggTaskGroup | 聚合任务 |
| Delivery Task | 数据投递任务 |
| Alert Rule | 告警规则 |
| Alert Template | 告警模板 |
| Alert History | 告警历史 |
| Notification Channel | 通知渠道 |
| Contact | 联系人 |
| Event Hub | 事件中心 |
| Metric Meta | 指标元数据 |
| ClusterCollector | 集群采集器 |
| NodeCollector | 节点采集器 |
| Cluster probe | 集群探针 |
| Metric drop | 指标废弃 |
| CMS resource tag | CMS 资源标签 |
| Grafana workspace | Grafana工作区 |
Metadata Query Mapping
| What You Need | How to Get It |
|---|---|
Metric business metadata (namespaces & product codes via meta namespaces; metric name, type, unit, dimensions via meta metrics) | meta namespaces / meta metrics |
| Prometheus labels, values & series inspection | metric promql labels / label-values / series |
Integration-module lookups (resource metadata, onboarding status, policy-scoped Kubernetes resources, Prometheus instance by policy) live in references/integration-common.md.
Module Routing
| User Intent Keywords | Commands | Module |
|---|---|---|
| onboarding, monitoring addon, policy, integration, addon release, integration resource, Kubernetes resource list, Namespace resources under policy, resources managed by policy, teardown, offboarding — common rules, load for every onboarding operation | integration integration resource | references/integration-common.md |
| container onboarding, ACK/ACS/ASI cluster onboarding, cluster fleet audit, which clusters are not onboarded | integration entity query | references/cs-onboarding.md (+ integration-common.md) |
| ECS host onboarding, ECS fleet audit, NodeCollector | integration entity query | references/ecs-onboarding.md (+ integration-common.md) |
| cloud service onboarding, RDS/SLB/ALB/Redis/MongoDB/PolarDB onboarding, cloud resource fleet audit | integration entity query | references/cloud-onboarding.md (+ integration-common.md) |
| batch cloud service metric onboarding, batch onboarding, cloud-batch-metrics | integration meta entity query | references/batch-onboarding-workflow.md (+ integration-common.md) |
integration policy diagnosis, health-check, troubleshoot policy, scrape config, ServiceMonitor/PodMonitor/custom collection diagnosis, custom-discover-<id> (a release name — read discoverType before treating it as a custom-job), job target | integration check-scrape-config integration job-target integration check-collector-target integration resource integration dashboard | references/integration-diagnosis.md |
| metric drop, drop metrics, dropMetrics, 指标废弃, 丢弃指标, 废弃指标, cluster probe metric-agent | integration collector integration addon-release | references/integration-management.md (+ integration-common.md) |
| add / change / remove CMS resource tags, 打标签, 修改标签, 删除标签, CMS resource tags, application service (APM/RUM) tags | tag | references/integration-management.md (+ integration-common.md) |
| Prometheus view, Prometheus aggregation view, create Prometheus aggregation view, Prometheus view create, Prometheus aggregation view diagnosis, Prometheus aggregation view health check, sub-instance status | prometheus view | references/prometheus-management.md |
| workspace, workspace create, workspace get, workspace list, workspace update, workspace delete | workspace | aliyun cms2 workspace --help |
| entity, entity query, CloudResource, EntityStore, cloud resource query, entity store query, resource metadata, instance details | entity query | aliyun cms2 entity --help |
| Prometheus instance, recordingRule, recording rule, AggTaskGroup | prometheus instance prometheus recording-rule | aliyun cms2 prometheus --help |
| meta, metric metadata, product code, meta-format | meta metrics meta namespaces | aliyun cms2 meta --help |
| metric, metric query, basic metrics, PromQL, promql query, label values, series | metric basic metric promql | aliyun cms2 metric --help |
| alert, rule, alert rule, alert template, alert history, patch, create rule, manage rule | alert rule alert template alert history | references/alerting.md |
| APM measureCode, group/filter/groupBy, baseUnit/displayUnit | alert rule (APM type) | references/apm-metrics.md |
| UModel metricSet, K8s pod metric, entity-based alert | alert rule (UModel type) | references/umodel-metrics.md |
| notification, contact, robot, webhook, notification recipients, dingTalk, bots, lark, weChat work | notification-channel contact notification-channel robot notification-channel webhook | references/alerting.md |
| event, event-hub, alert event, SLS event, incident | event-hub | references/event-hub.md |
| Grafana, Grafana workspace, managed Grafana instance, create/query/update/delete Grafana workspace | grafana workspace | aliyun cms2 grafana workspace --help |
| Grafana dashboard authoring, dashboard JSON, panel, PromQL panel, dashboard variables, data source placeholder | meta metrics metric promql integration storage | references/grafana-dashboard-rules.md |
| APM, application monitoring, agent install, Java agent, Golang agent, Python agent, Node.js agent, PHP agent, .NET agent, ack-onepilot, OpenTelemetry onboarding, K8s/ACK/ACS container onboarding, ECS host application onboarding, LicenseKey, proprietary agent, instgo, aliyun-bootstrap, probe setup, apm onboarding | apm service apm configuration | references/apm.md |
| AI observability, Dify, LangChain, LangGraph, DashScope, AgentScope, OpenAI, Coze, OpenClaw, CoPaw, Hermes, LLM monitoring, AI tracing, AI agent monitoring, custom instrumentation | apm service apm configuration integration addon | references/ai.md |
| RUM, Real User Monitoring, User Experience Monitoring, frontend monitoring, web monitoring, H5, mobile app monitoring, Android crash, iOS crash, JS error, page performance, miniapp monitoring, create RUM app, RUM SDK, pid, serviceId, endpoint | rum service rum configuration | references/rum.md |
| resource group query | resource-group | aliyun cms2 resource-group --help |
Commands not listed above — see aliyun cms2 --help.
FAQ
What does the alibabacloud-cms-manage skill do?
Entry skill for the aliyun CLI distribution of CloudMonitor (CMS). Use when the user mentions aliyun cms2, CloudMonitor, CMS commands, or any CMS module operation such as Integration Policy/Center, APM, RUM, Prometheus Service, Recording rule, alert rule, alert template, alert history, event hub, SLS event, PromQL, cloud resource, service observability, monitoring onboarding, metric query, etc.
How do I install alibabacloud-cms-manage?
Run npx skills add https://github.com/aliyun/alibabacloud-aiops-skills --skill alibabacloud-cms-manage in your project. The skills CLI detects the agents you use and copies the skill into the right folder. For Claude Code you can also place the skill folder at ~/.claude/skills/alibabacloud-cms-manage/ (personal) or .claude/skills/alibabacloud-cms-manage/ (project).
Which AI agents can use alibabacloud-cms-manage?
Any agent that supports the open Agent Skills format (a folder with a SKILL.md file), including Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Windsurf, Cline, OpenCode and more.
Is alibabacloud-cms-manage safe to install?
No partner security audit is available yet. Skills can run commands and scripts, so read the SKILL.md and any bundled files before installing.
Who maintains alibabacloud-cms-manage?
It is published by aliyun in aliyun/alibabacloud-aiops-skills. It has 245 recorded installs.
More from aliyun/alibabacloud-aiops-skills
alibabacloud-find-skills
aliyun/alibabacloud-aiops-skills
Use this skill when users want to search, discover, browse, or find Alibaba Cloud (阿里云) agent skills. Triggers include: "find a skill for X", "search alicloud skills", "阿里云有什么 skill","阿里云", "搜索阿里云技能", "有没有管理 ECS/RDS/OSS 的 skill", "阿里云 skills 有哪些类目", "帮我找一个 skill", "browse alicloud skills", "list alicloud skill categories", "is there an alicloud skill that can...", "what alicloud skills are available", "XX Skill 的内容是什么", "我想了解阿里云 XX Skill 具体做什么","帮我安装阿里云 Skill","使用阿里云相关的skill", "阿里云 agent skill 市场", "搜一下阿里云的 skill", "建一个数据分析项目有没有相关 skill".
alibabacloud-dms-skill
aliyun/alibabacloud-aiops-skills
Alibaba Cloud DMS Database Read/Write Skill. Use this skill to search for target databases in DMS and execute SQL queries and data modifications. Triggers: "DMS query", "database query", "execute SQL", "search database", "DMS SQL", "insert data", "update data".
alibabacloud-sls-query
aliyun/alibabacloud-aiops-skills
Alibaba Cloud SLS (Simple Log Service) log query & analysis skill. Use this skill to help users write, explain, optimize, execute, or troubleshoot SLS index search, SQL analytics, and SPL scan/pipeline statements through the aliyun CLI. Triggers: "SLS 查询", "SLS 分析", "日志查询", "日志分析", "log query", "analyze sls logs", "aliyun log query".
alibabacloud-workbench-cli
aliyun/alibabacloud-aiops-skills
Agent-native CLI for managing ECS instances without public IPs, primarily for single-instance operations. It supports millisecond-level remote command execution, large file transfers up to 1GB, and TCP port forwarding. It offers four authentication modes: AK, RamRoleArn, CredentialsCmd, and CredentialsURI. Use it to run commands, deploy code, view logs, check processes, transfer files, set up port forwarding, or query and filter ECS instance lists.
alibabacloud-ecs-diagnose
aliyun/alibabacloud-aiops-skills
Comprehensive Alibaba Cloud ECS instance diagnostics skill. Performs systematic troubleshooting including cloud platform status checks and GuestOS internal diagnostics via Cloud Assistant. Use when users report server connectivity issues, SSH timeout, instance lag, website unavailability, disk full, CPU/memory alerts, system event notifications, or abnormal instance status. Triggers: "ECS", "instance", "server", "cannot connect", "SSH", "timeout", "slow", "disk full", "network", "CPU high", "memory high", "status check", "system event", "diagnose", "troubleshoot", "disk performance", "disk IO slow", "IOPS", "IO latency"
alibabacloud-cli-guidance
aliyun/alibabacloud-aiops-skills
Guide users to manage Alibaba Cloud resources using the Aliyun CLI command-line tool. Covers CLI installation, credential configuration, plugin management, command construction, and error troubleshooting. Use this skill when the user wants to operate Alibaba Cloud services from the terminal — including ECS (云服务器, cloud servers), Function Compute (函数计算, serverless), RDS (云数据库, databases), OSS (对象存储, object storage), SLS (日志服务, log service), VPC (专有网络, networking), ESS (弹性伸缩, auto scaling), and any other Alibaba Cloud product. Also use when the user mentions "aliyun", "阿里云", "阿里云CLI", "命令行", asks about CLI plugin installation, encounters Aliyun CLI errors (InvalidAccessKeyId, SignatureDoesNotMatch, Throttling), or needs help constructing aliyun commands with correct parameter syntax.
Popular devops & cloud skills
microsoft-foundry
microsoft/azure-skills
Build, deploy, evaluate, optimize, fine-tune, and manage Microsoft Foundry agents, models, and resources end to end. USE FOR: foundry, azd ai agent, azd provision/deploy, hosted agent scaffold/develop/run/deploy/troubleshoot, prompt agent create, managed harness agent create, GitHub Copilot harness agent, create agent, update agent, add tool to agent, invoke agent, agent insights, evaluate agent, batch eval, continuous eval, continuous monitoring, agent CI/CD, optimize prompt, improve prompt, optimize agent instructions, Agent Optimizer scaffold, dataset curation from traces, deploy model, model fine-tuning (SFT/DPO/RFT), Foundry project, RBAC, role assignment, permissions, quota, capacity, region, deployment failure, AI Services, create Foundry resource, knowledge index, customize deployment, onboard, availability, training-data, grader, distillation, large file upload. DO NOT USE FOR: Azure Functions, App Service, general Azure deploy (use azure-deploy), general Azure prep (use azure
azure-diagnostics
microsoft/azure-skills
Debug Azure production issues on Azure using AppLens, Azure Monitor, resource health, and safe triage. WHEN: debug production issues, troubleshoot app service, app service high CPU, app service deployment failure, troubleshoot container apps, container app 503 or target port mismatch, troubleshoot functions, function telemetry missing or stopped after a config/connection-string change, troubleshoot AKS, VM RDP, Linux SSH, VM black screen, can't connect to VM, reset VM password, NSG or firewall blocking, kubectl cannot connect, kube-system/CoreDNS failures, pod pending, crashloop, node not ready, upgrade failures, analyze logs, KQL, insights, image pull failures, cold start issues, health probe failures, resource health, root cause of errors, troubleshoot event hubs, troubleshoot service bus, messaging SDK error, AMQP connection failure, message lock lost, lock renewal during long-running processing, service bus dead letter.
azure-prepare
microsoft/azure-skills
Prepare azd-based Azure projects for deployment: generates azure.yaml, infrastructure (Bicep/Terraform), and Dockerfiles for the Azure Developer CLI (azd) workflow. USE ONLY when the user explicitly wants to use azd as the deployment tool, or the project already has an azure.yaml file. DO NOT USE FOR: non-azd deployments, Python App Service code-only deploys (use python-appservice-deploy), or cross-cloud migration (use azure-cloud-migrate). WHEN: prepare app for azd, create azure.yaml, set up azd infrastructure, modernize app for Azure with azd, deploy with azd, function app, timer trigger, service bus trigger, event-driven function, managed identity, generate Bicep, generate Terraform, create and deploy to Azure.
azure-ai
microsoft/azure-skills
Use for Azure AI: Search, Speech, OpenAI, Document Intelligence. Helps with search, vector/hybrid search, speech-to-text, text-to-speech, transcription, OCR. WHEN: AI Search, query search, vector search, hybrid search, semantic search, speech-to-text, text-to-speech, transcribe, OCR, convert text to speech.
azure-deploy
microsoft/azure-skills
Execute Azure deployments for ALREADY-PREPARED applications that have existing .azure/deployment-plan.md and infrastructure files. DO NOT use this skill when the user asks to CREATE a new application — use azure-prepare instead. This skill runs azd up, azd deploy, terraform apply, and az deployment commands with built-in error recovery. Requires .azure/deployment-plan.md from azure-prepare and validated status from azure-validate. WHEN: "run azd up", "run azd deploy", "execute deployment", "push to production", "push to cloud", "go live", "ship it", "bicep deploy", "terraform apply", "publish to Azure", "launch on Azure". DO NOT USE WHEN: "create and deploy", "build and deploy", "create a new app", "set up infrastructure", "create and deploy to Azure using Terraform" — use azure-prepare for these.
azure-validate
microsoft/azure-skills
Pre-deployment validation for Azure readiness. Run deep checks on configuration, infrastructure (Bicep or Terraform), RBAC role assignments, managed identity permissions, and prerequisites before deploying. WHEN: validate my app, check deployment readiness, run preflight checks, verify configuration, check if ready to deploy, validate azure.yaml, validate Bicep, test before deploying, troubleshoot deployment errors, validate Azure Functions, validate function app, validate serverless deployment, verify RBAC roles, check role assignments, review managed identity permissions, what-if analysis, validate Container Apps deployment.